Loading…
This event has ended. Visit the official site or create your own event on Sched.

The Seventh Annual Tom Tom Founders Festival is a week-long celebration of innovators, visionaries, and artists who are shaping small cities. The Festival occurs at dozens of venues throughout downtown Charlottesville. 

Log in to bookmark your favorites and sync them to your phone or calendar.

Machine Learning [clear filter]
Thursday, April 12
 

9:00am

Machine Learning Coffee Hour
You need this ticket from Eventbrite to sign up:  Summit Badge - Advance Badge.
Start your day with some caffeine and conversation alongside Machine Learning innovators.

Sponsors
avatar for Capital One

Capital One

Capital One is a diversified bank that offers a broad array of financial products and service to consumers, small businesses and commercial clients. 
avatar for S&P Global

S&P Global

S&P Global Inc. (prior to April 2016 McGraw Hill Financial, Inc., and prior to 2013 McGraw Hill Companies) is an American publicly traded corporation headquartered in New York City. Its primary areas of business are financial information and analytics. It is the parent company of... Read More →


  • Festival (FREE), Included in Summit Badge, or Individually Ticketed Summit

9:45am

Applied Machine Learning Conference
Applied Machine Learning Conference speakers and programs now announced! Be sure to add individual sessions below to your Sched!

This second annual conference features researchers, entrepreneurs, and practitioners who are at the forefront of understanding and implementing big data and machine learning. This day of presentations and flash talks is focused on applied learning and deep science, with receptions and expos throughout the day.

Visit the AMLC homepage HERE and purchase your ticket HEREStudent tickets are availalble.

Sponsors
avatar for Capital One

Capital One

Capital One is a diversified bank that offers a broad array of financial products and service to consumers, small businesses and commercial clients. 
avatar for S & P Global

S & P Global

We provide intelligence that is essential for companies, governments and individuals to makedecisions with conviction.


10:15am

KEYNOTE: Deep Learning & Computer Graphics
Limited Capacity full
Adding this to your schedule will put you on the waitlist.
You need this ticket from Eventbrite to sign up:

Machine Learning Conference Badge

OR

Summit Badge - Advance Badge
Dr. David Luebke helped found NVIDIA Research in 2006 after eight years on the faculty of the University of Virginia. Today he run NVIDIA's research efforts at the intersection of computer graphics, machine learning, and virtual & augmented reality. Together with his colleagues David has written a book, created a major VR museum exhibit visited by over 110,000 people, taught an online course on parallel computing that has reached over 100,000 students, and authored over a hundred papers, articles, chapters, and patents. David is an NVIDIA Distinguished Inventor and a Fellow of the IEEE.

UPDATE: THERE WILL BE TWO OVERFLOW THEATERS LIVE STREAMING THE TALK

Speakers
avatar for David Luebke

David Luebke

VP Graphics Research, NVIDIA
Dr. David Luebke helped found NVIDIA Research in 2006 after eight years on the faculty of the University of Virginia. Today he runs NVIDIA's research efforts at the intersection of computer graphics, machine learning, and virtual & augmented reality. Together with his colleagues David... Read More →

Sponsors
avatar for Capital One

Capital One

Capital One is a diversified bank that offers a broad array of financial products and service to consumers, small businesses and commercial clients. 
avatar for S&P Global

S&P Global

S&P Global Inc. (prior to April 2016 McGraw Hill Financial, Inc., and prior to 2013 McGraw Hill Companies) is an American publicly traded corporation headquartered in New York City. Its primary areas of business are financial information and analytics. It is the parent company of... Read More →


11:00am

Velotonomy: beyond Uber, shared bikes and autonomous cars
In contrast to the industry’s dominant car-centric approach of concentrating on the 4-wheel automobile and its individual computation power, the PEV project focuses on the human-scale, pedestrian-friendly mobility solutions that bring more convenience, style and well-being to contemporary urban life.

Speakers
avatar for Phil Tinn

Phil Tinn

Research Scientist, MIT Media Lab - City Science Group
Phil Tinn, Research Scientist and co-lead of future mobility projects at MIT Media Lab, is a practitioner of innovation across scales and physical-digital media, with over 10 years experience in product, service, design, and organizational change strategies and execution... Read More →

Sponsors
avatar for Capital One

Capital One

Capital One is a diversified bank that offers a broad array of financial products and service to consumers, small businesses and commercial clients. 
avatar for S&P Global

S&P Global

S&P Global Inc. (prior to April 2016 McGraw Hill Financial, Inc., and prior to 2013 McGraw Hill Companies) is an American publicly traded corporation headquartered in New York City. Its primary areas of business are financial information and analytics. It is the parent company of... Read More →


  • Festival (FREE), Included in Summit Badge, or Individually Ticketed Summit

11:00am

Word Embeddings Under the Hood: How Neural Networks Learn from Language
Since their introduction in the early 2010s, word vector embedding models have exploded in popularity and use. They are one of the key breakthroughs that have enabled a new, state-of-the-art approach to natural language processing based on deep learning. But despite their impact, relatively few practitioners understand how word vector models work under the hood to capture the semantic relationships within natural language data and produce their remarkable results. Patrick Harrison opens up the black box of a popular word embedding algorithm and walks you through of how it works its magic. Along the way, Patrick also covers core neural network concepts, including hidden layers, loss gradients, backpropagation, and more. This talk is based on an excerpt from the forthcoming book Deep Learning with Text from O’Reilly Media.

Speakers
avatar for Patrick Harrison

Patrick Harrison

Director, AI Engineering, S&P Global
Patrick W. Harrison serves as Director of AI Engineering at S&P Global, a business and financial intelligence provider, and started the AI Engineering group there. Working in a major company for which data is both the primary raw material and finished product provides an unusually... Read More →

Sponsors
avatar for Capital One

Capital One

Capital One is a diversified bank that offers a broad array of financial products and service to consumers, small businesses and commercial clients. 
avatar for S&P Global

S&P Global

S&P Global Inc. (prior to April 2016 McGraw Hill Financial, Inc., and prior to 2013 McGraw Hill Companies) is an American publicly traded corporation headquartered in New York City. Its primary areas of business are financial information and analytics. It is the parent company of... Read More →


  • Festival (FREE), Included in Summit Badge, or Individually Ticketed Summit

11:05am

Use of Deep Learning for Environmental Enteropathy Biopsy Image Analysis
Limited Capacity seats available

A major challenge of interpreting clinical biopsy images to diagnose disease is the often striking overlap in histopathology between distinct but related conditions. There is a major clinical need to develop new methods in data science to allow clinicians to translate heterogeneous biomedical data extracted from patient samples into accurate, quantitative, and precise diagnostics. The development of such processes in high dimensional clinical research data will support the progress of “precision medicine”, with improved diagnostics, treatments, and clinical outcomes for patients. In this presentation we show preliminary results from an image analysis platform for the automated extraction of quantitative morphologic phenotypes from gastrointestinal (GI) biopsy images. This method will capture complex GI disease phenotypes, which cannot be measured directly using molecular approaches. We also aim to develop methods in data science to support the integration of this data with clinical and molecular data, enabling the construction of biologically informative and clinically useful integrative prognostic models in our case, in Environmental Enteropathy. EE is an acquired small intestinal inflammatory condition underlying high rates of linear growth stunting in children <5y of age in low- and middle-income countries. The histologic appearance of duodenal EE biopsies significantly overlaps with celiac enteropathy. To identify novel features that distinguish these two enteropathies from healthy tissue, we characterized intestinal mucosal alterations in Pakistani infants with EE using Convolutional Neural Networks (CNN). 

Speakers
avatar for Sana Syed

Sana Syed

Assistant Professor of Pediatrics, Division of Gastroenterology, Hepatology & Nutrition, University of Virginia
Sana Syed, MD MS is a Pediatric Gastroenterologist with advanced training in Nutrition at the University of Virginia. She is on faculty at the University of Virginia in the Center for Global Health with an adjunct faculty appointment at the Aga Khan University, Department of Pediatrics... Read More →

Sponsors
avatar for Capital One

Capital One

Capital One is a diversified bank that offers a broad array of financial products and service to consumers, small businesses and commercial clients. 
avatar for S&P Global

S&P Global

S&P Global Inc. (prior to April 2016 McGraw Hill Financial, Inc., and prior to 2013 McGraw Hill Companies) is an American publicly traded corporation headquartered in New York City. Its primary areas of business are financial information and analytics. It is the parent company of... Read More →


Thursday April 12, 2018 11:05am - 11:35am
Violet Crown - Theater 3 200 W Main St, Charlottesville, VA 22902

11:20am

A Modified K-Means Clustering Approach to Redrawing US Congressional Districts
Partisan gerrymandering threatens to undermine our democracy. It allows a political party in power to redraw district lines to its electoral advantage, making elections less competitive and less reflective of the will of the public. Most, including the US Supreme Court, agree that gerrymandering is an issue — but they stumble when asked to specify a more just policy for exactly how these lines should be drawn. How do you define an objective measure of political fairness?

Just as algorithms have made gerrymandering easier to implement and more widespread than ever, they can be used to reverse it. With congressional district lines set to be redrawn following the 2020 Census, this is the time to tackle this issue head on.

Speakers
avatar for Jack Prominski

Jack Prominski

Master's Student, UVA Data Science Institute
Jack Prominski is a Master’s Student in Data Science at the UVA Data Science Institute. His research centers around applying NLP techniques to scientific journal articles as part of a capstone project for the Public Library of Science. Prior to graduate school, Jack worked in e-commerce... Read More →

Sponsors
avatar for Capital One

Capital One

Capital One is a diversified bank that offers a broad array of financial products and service to consumers, small businesses and commercial clients. 
avatar for S&P Global

S&P Global

S&P Global Inc. (prior to April 2016 McGraw Hill Financial, Inc., and prior to 2013 McGraw Hill Companies) is an American publicly traded corporation headquartered in New York City. Its primary areas of business are financial information and analytics. It is the parent company of... Read More →


11:30am

Embed, Encode, Attend, Predict: A four-step framework for understanding neural network approaches to Natural Language Understanding problems
While there is a wide literature on developing neural networks for natural language understanding, the networks all have the same general architecture, determined by basic facts about the nature of linguistic input. In this talk I name and explain the four components (embed, encode, attend, predict), give a brief history of approaches to each subproblem, and explain two sophisticated networks in terms of this framework -- one for text classification, and another for textual entailment. The talk assumes a general knowledge of neural networks and machine learning. The talk should be especially suitable for people who have been working on computer vision or other problems. Just as computer vision models are designed around the fact that images are two or three-dimensional arrays of continuous values, NLP models are designed around the fact that text is a linear sequence of discrete symbols that form a hierarchical structure: letters are grouped into words, which are grouped into larger syntactic units (phrases, clauses, etc), which are grouped into larger discursive structures (utterances, paragraphs, sections, etc). Because the input symbols are discrete (letters, words, etc), the first step is "embed": map the discrete symbols into continuous vector representations. Because the input is a sequence, the second step is "encode": update the vector representation for each symbol given the surrounding context. You can't understand a sentence by looking up each word in the dictionary --- context matters. Because the input is hierarchical, sentences mean more than the sum of their parts. This motivates step three, attend: learn a further mapping from a variable-length matrix to a fixed-width vector, which we can then use to predict some specific information about the meaning of the text.

Speakers
avatar for Matthew Honnibal

Matthew Honnibal

Founder, Explosion AI
Matthew Honnibal is the creator and lead developer of spaCy, one of the most popular libraries for Natural Language Processing. He has been publishing research on NLP since 2005, with a focus on syntactic parsing and other structured prediction problems. He left academia to start... Read More →

Sponsors
avatar for Capital One

Capital One

Capital One is a diversified bank that offers a broad array of financial products and service to consumers, small businesses and commercial clients. 
avatar for S&P Global

S&P Global

S&P Global Inc. (prior to April 2016 McGraw Hill Financial, Inc., and prior to 2013 McGraw Hill Companies) is an American publicly traded corporation headquartered in New York City. Its primary areas of business are financial information and analytics. It is the parent company of... Read More →


11:40am

Decrypting Crypto Fascist Symbolism With Tensorflow
Last year the events of the Unite The Right Rally here in Charlottesville made a tremendous impact on the public's awareness and understanding of modern white supremacy and its propagation. In this talk, Emily will discuss how she used machine learning to identify white supremacist symbolism and how you can create your own custom machine learning model.

Speakers
avatar for Emily Crose

Emily Crose

Threat Hunter, Ironnet Cybersecurity
Emily Crose is a security researcher and professional with over 10 years of experience including a total of 7 between being an officer for the Central Intelligence Agency and the National Security Agency. She is currently directing the Nemesis project in her free time and currently... Read More →

Sponsors
avatar for Capital One

Capital One

Capital One is a diversified bank that offers a broad array of financial products and service to consumers, small businesses and commercial clients. 
avatar for S&P Global

S&P Global

S&P Global Inc. (prior to April 2016 McGraw Hill Financial, Inc., and prior to 2013 McGraw Hill Companies) is an American publicly traded corporation headquartered in New York City. Its primary areas of business are financial information and analytics. It is the parent company of... Read More →


11:40am

Joining Separate Paradigms: Natural Language Processing and Deep Neural Networks to Characterize Neuronal Cell Type and Function
Limited Capacity full
Adding this to your schedule will put you on the waitlist.

Revolutionary progress in technology has opened new biological questions about cell types, states, and gene regulation on a single-cell scale. However, with the ability to understand how an individual cell functions discretely comes new data-oriented questions about dimensionality, feature reduction, and novel methods to account for noise. The central objective of our research is to refine novel applications for single-cell data to better characterize sub-populations of cell types in various regions of the brain. 

Hold onto your seats because this presentation will take you on a journey from the cells in your brain to the underpinnings of data science in biomedicine!

Speakers
avatar for Cait Dreisbach

Cait Dreisbach

PhD Candidate, University of Virginia
Cait Dreisbach is a PhD Candidate at the University of Virginia School of Nursing. She is a recent alumna of the Masters in Data Science program at the UVA Data Science Institute. Her research interests are in microbial genomics and women's health.
avatar for Morgan Wall

Morgan Wall

Master's Student, UVA Data Science Institute
Morgan Wall graduated from the University of Virginia in Charlottesville, Virginia in May 2015 with a dual Bachelor of Arts in Biology and Economics. Following graduation, she worked in the field of Melanoma research at the University of Virginia Human Immune Therapy Center, investigating... Read More →
avatar for Ali Zaidi

Ali Zaidi

Master's Student, UVA Data Science Institute
Ali Zaidi achieved a Bachelor’s in Neuroscience and Information Technology from George Mason University in Northern Virginia in May of 2017. During his time at George Mason University, he was involved in multiple research opportunities within the Krasnow Institute of Advanced Study... Read More →

Sponsors
avatar for Capital One

Capital One

Capital One is a diversified bank that offers a broad array of financial products and service to consumers, small businesses and commercial clients. 
avatar for S&P Global

S&P Global

S&P Global Inc. (prior to April 2016 McGraw Hill Financial, Inc., and prior to 2013 McGraw Hill Companies) is an American publicly traded corporation headquartered in New York City. Its primary areas of business are financial information and analytics. It is the parent company of... Read More →


Thursday April 12, 2018 11:40am - 11:55am
Violet Crown - Theater 3 200 W Main St, Charlottesville, VA 22902

12:00pm

Building an Applied Machine Learning Laboratory for Undergraduate Research and Education
This presentation will detail the creation of the Data Science and Applied Machine Learning Laboratory and complementary Data Science curriculum for undergraduate research and education at James Madison University. The current state of the lab supports research, development and applications in four key areas: (i) small-scale autonomous vehicles, (ii) image classification and object detection, (iii) algorithm development for single-layer and multi-layer artificial neural networks, and (iv) introduction to deep neural networks (deep learning) with parallel distributed computing. The research focus areas of the lab are being designed to emphasize core underlying principles, techniques and methodologies specific to applied machine learning and data science. The curriculum development parallels this approach with courses that provide students with many of the foundational, translational, professional and ethical skills that will make it possible for them to become effective practitioners of data science and machine learning. Both research and pedagogy use two production-ready Python frameworks: Anaconda’s Data Science Platform and Google’s TensorFlow. It is envisioned that this integrated research and pedagogical approach may become a “best practice” that informs the development and assessment of new and emerging machine learning and data science programs at other institutions.

Speakers
avatar for Anthony Teate

Anthony Teate

Professor, James Madison University
Dr. Anthony A. Teate is a tenured Professor in the School of Integrated Sciences located in the College of Integrated Science and Engineering at James Madison University, where he is also the Director of the Data Science and Applied Machine Learning Laboratory. His current areas of... Read More →

Sponsors
avatar for Capital One

Capital One

Capital One is a diversified bank that offers a broad array of financial products and service to consumers, small businesses and commercial clients. 
avatar for S&P Global

S&P Global

S&P Global Inc. (prior to April 2016 McGraw Hill Financial, Inc., and prior to 2013 McGraw Hill Companies) is an American publicly traded corporation headquartered in New York City. Its primary areas of business are financial information and analytics. It is the parent company of... Read More →


12:00pm

How I Learned to Stop Worrying and Learned to Love Graph Databases
Limited Capacity full
Adding this to your schedule will put you on the waitlist.

In the age of big data context matters. Graph databases encode context and relationships into the data. We used that context to look at things like how infection spread in the hospital, how patients move through the hospital, and how teams interact. This talk will explain what a graph database is, how to get started, and how to use social network analysis to get deep insights or make predictions.

Speakers
avatar for Michael Zelenetz

Michael Zelenetz

Data Insights Programmer, New York Presbyterian hospital
Michael works in Analytics at New York Presbyterian—one of the country’s foremost academic medical centers. When not working he can he found running or climbing with his two sons.

Sponsors
avatar for Capital One

Capital One

Capital One is a diversified bank that offers a broad array of financial products and service to consumers, small businesses and commercial clients. 
avatar for S&P Global

S&P Global

S&P Global Inc. (prior to April 2016 McGraw Hill Financial, Inc., and prior to 2013 McGraw Hill Companies) is an American publicly traded corporation headquartered in New York City. Its primary areas of business are financial information and analytics. It is the parent company of... Read More →


Thursday April 12, 2018 12:00pm - 12:15pm
Violet Crown - Theater 3 200 W Main St, Charlottesville, VA 22902

12:00pm

Rapid NLP Annotation Through Binary Decisions, Pattern Bootstrapping and Active Learning Using Prodigy and spaCy
Limited Capacity full
Adding this to your schedule will put you on the waitlist.

In this talk, I'll present a fast, flexible and even somewhat fun approach to named entity annotation. Using our approach, a model can be trained for a new entity type in only a few hours, starting from only a feed of unannotated text and a handful of seed terms. Given the seed terms, we first perform an interactive lexical learning phase, using a semantic similarity model that can be trained from raw text via an algorithm such as word2vec. The similarity model can be made to learn vectors for longer phrases by pre-processing the text, and abstract patterns can be created referencing attributes such as part-of-speech tags. The patterns file is then used to present the annotator with a sequence of candidate phrases, so that the annotation can be conducted as a binary choice. The annotator's eyes remain fixed near the centre of the screen, decisions can be made with a click, swipe or single keypress, and tasks are buffered to prevent delays. Using this interface, annotation rates of 10-30 decisions per minute are common. If the decisions are especially easy (e.g. confirming that instances of a phrase are all valid entities), the rate may be several times faster. As the annotator accepts or rejects the suggested phrases, the responses are used to start training a statistical model. Predictions from the statistical model are then mixed into the annotation queue. Despite the sparsity of the signal (binary answers on one phrase per sentence), the model begins to learn surprisingly quickly. A global neural network model is used, with beam-search to allow a form of noise-contrastive estimation training. The pattern matcher and entity recognition model is available in our open-source library spaCy, while the interface, task queue and workflow management are implemented in our annotation tool Prodigy.

Speakers
avatar for Ines Montani

Ines Montani

Founder, Explosion AI
Ines is a developer specializing in applications for AI technology. She's the co-founder of Explosion AI and a core developer of spaCy, the leading open-source library for Natural Language Processing in Python and Prodigy, an annotation tool for radically efficient machine teachi... Read More →

Sponsors
avatar for Capital One

Capital One

Capital One is a diversified bank that offers a broad array of financial products and service to consumers, small businesses and commercial clients. 
avatar for S&P Global

S&P Global

S&P Global Inc. (prior to April 2016 McGraw Hill Financial, Inc., and prior to 2013 McGraw Hill Companies) is an American publicly traded corporation headquartered in New York City. Its primary areas of business are financial information and analytics. It is the parent company of... Read More →


1:00pm

Real-time, Streaming Anomaly Detection for Cybersecurity with Redis-ML
Limited Capacity full
Adding this to your schedule will put you on the waitlist.

Network-based threats to personal information, company secrets and critical infrastructure are increasing in frequency and sophistication. A rapid response is critical for stopping these threats, but, with network speeds reaching 100Gb/s and beyond, network defenders are swamped by overwhelming amounts of traffic, hindering both detection and response. Data scientists maintain that machine learning has the potential to help fight these threats, but the current dominant “Big Data” strategy of backhauling data to a central repository, using tools such as Hadoop and Spark, adds significant latency while incurring additional costs for bandwidth, storage, and computational resources. Furthermore, “Big Data” platforms such as Spark are optimized primarily for data management and model training, not serving trained models in deployment. We will walk through the process of translating a multi-dimensional anomaly detection algorithm that operates in batch into a streaming algorithm suitable for deployment in a network sensor. Then we demonstrate how deploying these models with Redis and the Redis ML module leads to dramatic reduction in processing time, leading to the faster threat detection and response.

Speakers
avatar for Andrew Fast

Andrew Fast

Chief Data Scientist, CounterFlow AI, Inc
Andrew Fast is the Chief Data Scientist and co-founder of CounterFlow AI, where he leads the implementation of streaming machine learning algorithms for CounterFlow's next-generation network forensics platform, ThreatEye.Previously, Dr. Fast served as the Chief Scientist at Elder... Read More →

Sponsors
avatar for Capital One

Capital One

Capital One is a diversified bank that offers a broad array of financial products and service to consumers, small businesses and commercial clients. 
avatar for S&P Global

S&P Global

S&P Global Inc. (prior to April 2016 McGraw Hill Financial, Inc., and prior to 2013 McGraw Hill Companies) is an American publicly traded corporation headquartered in New York City. Its primary areas of business are financial information and analytics. It is the parent company of... Read More →


1:05pm

Can a Machine Be Racist or Sexist?
Limited Capacity seats available

Machine learning algorithms are coded procedures - a series of logical steps that use mathematical functions to find patterns in datasets - so is it possible for a machine learning model to be socially biased? How can social bias make its way into our data, our trained models, and our artificial intelligence, and how could that affect people in our society? What do we need to be aware of as data scientists in order to prevent our models from treating people unfairly and causing harm?

Speakers
avatar for Renee Teate

Renee Teate

Data Scientist, HelioCampus
Renee M. P. Teate is a Data Scientist at HelioCampus, and the creator of the Becoming a Data Scientist Podcast and @becomingdatasci twitter account. She has worked with data for her entire career - designing relational databases, creating reports and analyses, and most recently developing... Read More →

Sponsors
avatar for Capital One

Capital One

Capital One is a diversified bank that offers a broad array of financial products and service to consumers, small businesses and commercial clients. 
avatar for S&P Global

S&P Global

S&P Global Inc. (prior to April 2016 McGraw Hill Financial, Inc., and prior to 2013 McGraw Hill Companies) is an American publicly traded corporation headquartered in New York City. Its primary areas of business are financial information and analytics. It is the parent company of... Read More →


1:05pm

Two Years of Bayesian Bandits for E-Commerce
Limited Capacity filling up

At Monetate, we’ve deployed Bayesian bandits (both noncontextual and contextual) to help our clients optimize their e-commerce sites since early 2016. This talk is an overview of the lessons we’ve learned from both the processes of deploying real-time Bayesian machine learning systems at scale and building a data product on top of these systems that is accessible to non-technical users (marketers). This talk will cover:
  • The place of multi-armed bandits in the A/B testing industry,
  • Thompson sampling and the basic theory of Bayesian bandits,
  • Bayesian approaches for accommodating nonstationarity in bandit feedback,
  • User experience challenges in driving adoption of these technologies by nontechnical marketers.
We will focus primarily on noncontextual bandits and give a brief overview of these problems in the contextual setting as time permits.

Speakers
avatar for Austin Rochford

Austin Rochford

Chief Data Scientist, Monetate
Austin Rochford is Principal Data Scientist and Director of Monetate Labs. He is a founding member of Monetate Labs, where he does research and development for machine learning-driven marketing products. He is a recovering mathematician, a passionate Bayesian, and a PyMC3 developer... Read More →

Sponsors
avatar for Capital One

Capital One

Capital One is a diversified bank that offers a broad array of financial products and service to consumers, small businesses and commercial clients. 
avatar for S&P Global

S&P Global

S&P Global Inc. (prior to April 2016 McGraw Hill Financial, Inc., and prior to 2013 McGraw Hill Companies) is an American publicly traded corporation headquartered in New York City. Its primary areas of business are financial information and analytics. It is the parent company of... Read More →


Thursday April 12, 2018 1:05pm - 1:35pm
Violet Crown - Theater 3 200 W Main St, Charlottesville, VA 22902

1:20pm

The Art and Science of Bieberfinding
Limited Capacity full
Adding this to your schedule will put you on the waitlist.

Social media offers a diverse and intermeshed landscape of creators and consumers that cuts across traditional archetypes of genre and stardom. Given the speed of evolution, fractured landscape of delivery platforms, and sheer breadth of choices, discovering emerging talent becomes a now-classic big-data classification problem against a broad but shallow signal. Access to data remains the primary challenge to effective feature engineering in this space, and therefore to identifying the next Justin Bieber, political star, or terrorist leader. A cascade of models and feedback loops can address this problem by first scoring against a weak but suitably general source of information, such as Twitter, and then leveraging this rough set to prioritize subsequent, more-focused data collection. Once the search space has been sufficiently reduced, a deep learning (embedding) model can feasibly employed for similarity search and general inference.

Speakers
avatar for Alec Gosse

Alec Gosse

Lead Data Scientist, Metis Machine
Alec Gosse comes to data science from a study of how infrastructure investment and mobility patterns interact to drive environmental impacts. Prior to joining Metis Machine as Lead Data Scientist, Gosse worked to build unsupervised place understanding by fusing social media, GIS... Read More →

Sponsors
avatar for Capital One

Capital One

Capital One is a diversified bank that offers a broad array of financial products and service to consumers, small businesses and commercial clients. 
avatar for S&P Global

S&P Global

S&P Global Inc. (prior to April 2016 McGraw Hill Financial, Inc., and prior to 2013 McGraw Hill Companies) is an American publicly traded corporation headquartered in New York City. Its primary areas of business are financial information and analytics. It is the parent company of... Read More →


1:30pm

Social Bias in Machine Learning
Limited Capacity filling up

Machine learning algorithms are coded procedures - a series of logical steps that use mathematical functions to find patterns in datasets - so is it possible for a machine learning model to be socially biased? How can social bias make its way into our data, our trained models, and our artificial intelligence, and how could that affect people in our society? What do we need to be aware of as data scientists in order to prevent our models from treating people unfairly and causing harm?

Moderators
avatar for Renee Teate

Renee Teate

Data Scientist, HelioCampus
Renee M. P. Teate is a Data Scientist at HelioCampus, and the creator of the Becoming a Data Scientist Podcast and @becomingdatasci twitter account. She has worked with data for her entire career - designing relational databases, creating reports and analyses, and most recently developing... Read More →

Speakers
avatar for Emily Crose

Emily Crose

Threat Hunter, Ironnet Cybersecurity
Emily Crose is a security researcher and professional with over 10 years of experience including a total of 7 between being an officer for the Central Intelligence Agency and the National Security Agency. She is currently directing the Nemesis project in her free time and currently... Read More →
avatar for Natalie Evans Harris

Natalie Evans Harris

Chief Operating Officer, BrightHive
Both as Founder of Harris Data Consulting and COO of BrightHive, Ms. Evans Harris has dedicated over 16 years to driving the strategic use of data to answer some of our nation’s toughest questions and driving organizational success; 
avatar for Vicente Ordonez

Vicente Ordonez

Assistant Professor, University of Virginia
Vicente Ordonez is Assistant Professor in the Department of Computer Science at the University of Virginia. His research interests lie at the intersection of computer vision, natural language processing and machine learning. He is a recipient of an IBM Faculty Award and a Google Faculty... Read More →

Sponsors
avatar for Capital One

Capital One

Capital One is a diversified bank that offers a broad array of financial products and service to consumers, small businesses and commercial clients. 
avatar for S&P Global

S&P Global

S&P Global Inc. (prior to April 2016 McGraw Hill Financial, Inc., and prior to 2013 McGraw Hill Companies) is an American publicly traded corporation headquartered in New York City. Its primary areas of business are financial information and analytics. It is the parent company of... Read More →


1:40pm

Empowering People and Thwarting Machines
Limited Capacity full
Adding this to your schedule will put you on the waitlist.

Under the current data paradigm, third parties often capture, analyze, and make use of an individual’s data without that individual’s knowledge or consent. The uses of this data are often opaque, and even when an individual signs a “Terms of Service Agreement” it is questionable whether they truly provide informed consent. With the EU’s General Data Protection Regulation (GDPR), we’re seeing momentum to provide individuals with greater data privacy protection. As part of the Assembly on artificial intelligence and governance -- a joint project between the Berkman Klein Center at Harvard University and the MIT Media Lab -- we are creating technical and policy mechanisms that empower individuals to thwart third parties from successfully analyzing their data and to communicate that they do not want their data to be used in ways that they did not consent to. We are building a tool that directly embeds “Do Not Track”-like signature in images while simultaneously duping image classification systems.

Speakers
avatar for Thom Miano

Thom Miano

Data Scientist, RTI International
Thom Miano is a data scientist in the Center for Data Science at RTI International. Thom studies Artificial Intelligence as a multifaceted discipline, investigating the development of algorithmic architectures, the practical application of machine learning in health, education, and... Read More →

Sponsors
avatar for Capital One

Capital One

Capital One is a diversified bank that offers a broad array of financial products and service to consumers, small businesses and commercial clients. 
avatar for S&P Global

S&P Global

S&P Global Inc. (prior to April 2016 McGraw Hill Financial, Inc., and prior to 2013 McGraw Hill Companies) is an American publicly traded corporation headquartered in New York City. Its primary areas of business are financial information and analytics. It is the parent company of... Read More →


2:05pm

KEYNOTE: How to Fail Spectacularly at Analytics
Limited Capacity full
Adding this to your schedule will put you on the waitlist.
You need this ticket from Eventbrite to sign up:

Machine Learning Conference Badge

OR

Summit Badge - Advance Badge
Success stories abound that extoll the use of “analytics” in a wide variety of fields and on a wide swath of problems. But there are valuable lessons to be learned from what didn’t work, whether a spectacular failure or just bumps in the road. In this talk, we describe real-world examples, some anonymous and most from the speakers’ decades of analytical consulting. We will describe pitfalls, misconceptions, and uncertainties commonly encountered by companies in their path towards adoption of analytics that led to failure, lost time, and wasted investment.

UPDATE: THIS TALK WILL BE LIVESTREAMED INTO AN ADDITIONAL THEATER

Speakers
avatar for Antonia de Medinaceli

Antonia de Medinaceli

Co-Founder, Augmented
Antonia has spent the past two decades in all aspects of the data science industry. Starting out as a modeler, she applied machine learning algorithms to fraud detection, credit scoring, lead generation, and crime prediction problems among many others. She has worked with both commercial... Read More →

Sponsors
avatar for Capital One

Capital One

Capital One is a diversified bank that offers a broad array of financial products and service to consumers, small businesses and commercial clients. 
avatar for S&P Global

S&P Global

S&P Global Inc. (prior to April 2016 McGraw Hill Financial, Inc., and prior to 2013 McGraw Hill Companies) is an American publicly traded corporation headquartered in New York City. Its primary areas of business are financial information and analytics. It is the parent company of... Read More →


2:45pm

Explainable AI: Key Techniques and Societal Implications
Limited Capacity filling up

The field of Explainable AI strives to understand the inner workings of “black-box” algorithms to ensure their safe adoption in society. As advances in AI and, more specifically, machine learning, create more complex models, there is a growing regulatory and ethical appetite to understand the decision-making process of these models. This talk dives into state of the art methods (e.g., LIME and Integrated Gradients) for explaining deep learning models and considers the challenges inherent to this research field.

 


Speakers
avatar for Mark Ibrahim

Mark Ibrahim

Machine Learning Engineer, Capital One
Mark Ibrahim is a machine learning engineer at Capital One's Center for Machine Learning.
avatar for Ceena Modarres

Ceena Modarres

Data Scientist, Capital One
Ceena Modarres is a Data Scientist and Machine Learning Engineer at Capital One. He works on ML interpretability research and natural language processing applications within the organization. He has a M.S. in Reliability Engineering and lives in Manhattan's Lower East Side.

Sponsors
avatar for Capital One

Capital One

Capital One is a diversified bank that offers a broad array of financial products and service to consumers, small businesses and commercial clients. 
avatar for S&P Global

S&P Global

S&P Global Inc. (prior to April 2016 McGraw Hill Financial, Inc., and prior to 2013 McGraw Hill Companies) is an American publicly traded corporation headquartered in New York City. Its primary areas of business are financial information and analytics. It is the parent company of... Read More →


2:45pm

In Man's Image: Autonomy through Imitation Learning
Limited Capacity filling up

When we dream about artificial intelligence, I conjecture that most of us do not imagine random forests being applied to business intelligence. The dream is artificial general intelligence (AGI)! A physical agent that can emulate the human mind. Unfortunately, the current belief is that AGI will not be seen until the mid-century (assuming the Earth and our economies are still functioning at that time.) So what can we do now? We can keep working to build AIs capable of autonomy in smaller tasks (such as sorting blocks or driving a car). Autonomous systems are a natural stepping stone to achieving AGI. If we cannot create systems that are capable of successfully performing our trivial tasks, how can we build an intelligence comparable to the human intellect? In that same vein, why don't we teach autonomy by our own example? What is more obvious than to teach an agent independence than by imitating the experts (i.e. us)?

This talk provides an introduction to imitation learning, a training technique that utilizes expert demonstrations to teach AI agents how to perform particular tasks. It begins with a brief explanation of reinforcement learning and its relationship to imitation learning for autonomous agents.  Then, there will be a discussion on popular imitation learning techniques and algorithms (such as behavioral cloning, DAgger, and one-shot imitation learning) and their benefits and tradeoffs. Using CARLA, a high-fidelity, open-source simulator for autonomous vehicle research, this presentation will provide examples of how well the different techniques actually work in practice. Lastly, it will wrap up with a discussion about why imitation learning does not solve the problem of autonomy, and what still needs to be done.

Speakers
avatar for Bryce Freshcorn

Bryce Freshcorn

Senior AI Engineer, The MITRE Corp.
Bryce Freshcorn is an artificial intelligence engineer at the MITRE Corp. His current research interests are focused on computer vision applications for end-to-end modeling in autonomous systems and healthcare diagnostics. His other passion (born from his experiences in the defense... Read More →

Sponsors
avatar for Capital One

Capital One

Capital One is a diversified bank that offers a broad array of financial products and service to consumers, small businesses and commercial clients. 
avatar for S&P Global

S&P Global

S&P Global Inc. (prior to April 2016 McGraw Hill Financial, Inc., and prior to 2013 McGraw Hill Companies) is an American publicly traded corporation headquartered in New York City. Its primary areas of business are financial information and analytics. It is the parent company of... Read More →


Thursday April 12, 2018 2:45pm - 3:15pm
Violet Crown - Theater 3 200 W Main St, Charlottesville, VA 22902

2:45pm

The Florist, the GPU, and Deep Learning (DL/Computer Vision Intro)
Dr. May Casterline is a data scientist/image scientist/software developer with a background in satellite and airborne imaging systems. Her research interests include deep learning, hyperspectral and multispectral imaging, innovative applications of machine learning approaches to remote sensing data, and creative software solutions to challenging geospatial problems. She holds a PhD and Bachelors of Science in Imaging Science from Rochester Institute of Technology, with a focus on remote sensing. In industry she has acted as a product owner, technical lead, lead developer, and image scientist on both research initiatives and development projects. As a Senior Solutions Architect at NVIDIA, Dr. Casterline works with both industry and government to help enable developers, engineers, data scientists and analysts integrate artificial intelligence and GPU-accelerated solutions into their workflows and products.

Speakers
avatar for May Casterline

May Casterline

Senior Solutions Architect, NVIDIA
Dr. May Casterline is a data scientist/image scientist/software developer with a background in satellite and airborne imaging systems. Her research interests include deep learning, hyperspectral and multispectral imaging, innovative applications of machine learning approaches to remote... Read More →

Sponsors
avatar for Capital One

Capital One

Capital One is a diversified bank that offers a broad array of financial products and service to consumers, small businesses and commercial clients. 
avatar for S&P Global

S&P Global

S&P Global Inc. (prior to April 2016 McGraw Hill Financial, Inc., and prior to 2013 McGraw Hill Companies) is an American publicly traded corporation headquartered in New York City. Its primary areas of business are financial information and analytics. It is the parent company of... Read More →


3:25pm

Training Drone Image Models with Grand Theft Auto
Limited Capacity full
Adding this to your schedule will put you on the waitlist.

Computer vision models often require large amounts of labeled and domain-relevant data for effective training. For analyzing drone video, models trained on existing computer vision datasets such as ImageNet don’t always transfer to the desired domain well. CCRi Data Scientist Monica Rajendiran will describe how she and her team take advantage of the photo-realism and access to high-level semantic information in the video game Grand Theft Auto to train computer vision models that can perform object detection and image captioning on real drone, security camera, and even infrared footage in the face of poor video quality, compression artifacts, and clutter.

Speakers
avatar for Monica Rajendiran

Monica Rajendiran

Data Scientist, CCRi
Monica Rajendiran is a Data Scientist at CCRi, where she researches and develops deep learning techniques for machine aided object detection and text generation in full motion video. When not working on personal machine learning projects, she’s running Charlottesville’s many hills... Read More →

Sponsors
avatar for Capital One

Capital One

Capital One is a diversified bank that offers a broad array of financial products and service to consumers, small businesses and commercial clients. 
avatar for S&P Global

S&P Global

S&P Global Inc. (prior to April 2016 McGraw Hill Financial, Inc., and prior to 2013 McGraw Hill Companies) is an American publicly traded corporation headquartered in New York City. Its primary areas of business are financial information and analytics. It is the parent company of... Read More →


3:25pm

Harnessing Machine Learning to Enable Discovery
Limited Capacity full
Adding this to your schedule will put you on the waitlist.

There has been an explosion in the number of commercially available satellite images produced every day. The democratization of space technology has catalyzed a revolution of the commercial space industry which is now rapidly transforming from one company imaging 5 million km2 a day to up to 10 companies imaging 200 million km2 a day, and from constellations of a handful of satellites to constellations of hundreds of satellites. Today, the number of images being generated by this rapidly evolving commercial space industry far exceeds human scales. This talk will explore how we can harness machine learning to enable global discovery at scale. Fueled by advancements in machine learning algorithms, GPUs, and the availability of labeled datasets, we have dramatically improved our ability to extract insight from this explosion of data to understand changes across the globe.

Speakers
avatar for Mikel Rodriguez

Mikel Rodriguez

Director of Computer Vision Research, MITRE
Mikel Rodriguez leads computer vision research at MITRE, a federally funded research and development center. For the past fifteen years his main research passion and interest has been the exploration of how we can use AI and in particular computer vision to help us solve problems... Read More →

Sponsors
avatar for Capital One

Capital One

Capital One is a diversified bank that offers a broad array of financial products and service to consumers, small businesses and commercial clients. 
avatar for S&P Global

S&P Global

S&P Global Inc. (prior to April 2016 McGraw Hill Financial, Inc., and prior to 2013 McGraw Hill Companies) is an American publicly traded corporation headquartered in New York City. Its primary areas of business are financial information and analytics. It is the parent company of... Read More →


Thursday April 12, 2018 3:25pm - 3:55pm
Violet Crown - Theater 3 200 W Main St, Charlottesville, VA 22902

3:30pm

News Article Classification for Improving Estimates of Arrest-Related Deaths in the US
Limited Capacity seats available

National statistics of police-involved deaths have been hampered by databases that rely on voluntary reporting by a decentralized system of more than 18,000 independent law enforcement agencies. As part of a redesign of the Arrest Related Deaths (ARD) program by BJS, RTI developed a coding and classification pipeline that applies machine learning techniques to identify deaths from open information sources, including news articles and official reports from state and local law enforcement agencies. This hybrid approach led to an annual estimate of 1,900 arrest-related deaths in the U.S., and 1,200 law enforcement homicides, in line with estimates from the 2015 capture recapture analysis and other independent media sources (Banks, Ruddle, Kennedy & Planty, 2016). The pipeline created in order to identify news articles representing arrested related deaths is comprised of several sequential computational processes. The initial step of the pipeline collects news articles from media monitoring services, where the articles collected have a title or text that contains a word from a list of domain relevant keywords. Following this step, the articles are deduplicated to remove articles that have significant overlap in article title or text. The remaining articles are run through a relevancy classifier - a machine learning model that predicts whether the article describes an arrest related death. A human coder makes the final determination of whether an article meets our definition of an arrest related death, and then links multiple articles to a specific decedent, if necessary, through a web interface. All told, this machine learning based system typically results in an 87% reduction in the total volume of articles. This project, reported by BJS, has been featured in numerous media outlets, including The Guardian, fivethirtyeight.com (who also included it among the Best Data Stories of 2016), and The Measure of Everyday Life podcast.

Speakers
avatar for Peter Baumgartner

Peter Baumgartner

Data Scientist, RTI International
Peter Baumgartner is a data scientist at RTI International, a non-profit research institute. He applies natural language processing, machine learning, and design thinking to build things that help people.


3:30pm

KEYNOTE: Defending the Digital Space: Fake News, Hate Speech, and Big Brother
Limited Capacity seats available

The internet's core strengths—its freedom and openness—increasingly seem to be double-edged swords, undermining trust and available for exploitation for criminals and companies alike. Can we safeguard our society and privacy online?

Moderators
avatar for Siva Vaidhyanathan

Siva Vaidhyanathan

Robertson Professor of Media Studies, Director, Center for Media and Citizenship at the University of Virginia
Siva Vaidhyanathan is the Robertson Professor of Media Studies and director of the Center for Media and Citizenship at the University of Virginia.He is the author of Antisocial Media: How Facebook Disconnects Us and Undermines Democracy (Oxford University Press, 2018). He also wrote... Read More →

Speakers
avatar for Renee Diresta

Renee Diresta

Data for Democracy, Head of Policy
DiResta is a founding advisor to the Center for Humane Technology, Head of Policy at Data for Democracy, and an advocate for vaccination in California. She is a cofounder of Haven, a private marketplace for booking ocean freight shipments. Previously, Renee was a principal at seed-stage... Read More →
avatar for Garrett Graff

Garrett Graff

Executive Director, Cybersecurity and Technology Program
Graff is a distinguished magazine journalist and historian, spending more than a dozen years covering politics, technology, and national security. Today, he serves as the director of the Aspen Institute’s cybersecurity and technology program, and is a contributor to WIRED, Longreads... Read More →
avatar for Cathy O'Neil

Cathy O'Neil

Author, Weapons of Math Destruction
Cathy O’Neil is a regular contributor to Bloomberg View and wrote the book "Weapons of Math Destruction: How Big Data Increases Inequality and Threatens Democracy." She earned a Ph.D. in math from Harvard, was a postdoc at the MIT math department, and a professor at Barnard College... Read More →

Sponsors
avatar for C-Ville Weekly

C-Ville Weekly

C-VILLE Weekly is a free newspaper with a weekly circulation of 23,000 copies distributed in Charlottesville and its surrounding counties. In addition, we publish seven magazines throughout the year: Abode, a monthly home and landscape magazine; C-BIZ, a quarterly business publication... Read More →
avatar for University of Virginia College and Graduate School of Arts & Sciences

University of Virginia College and Graduate School of Arts & Sciences

The University of Virginia College and Graduate School of Arts & Sciencesis the largest of the University of Virginia's ten schools. Consisting of both a graduate and an undergraduate program, the College comprises the liberal artsand humanities section of the University... Read More →


3:55pm

Learning From Billions of News Reading Sessions
Limited Capacity seats available

What can we learn from a one-billion-person live-poll of the Internet? Parse.ly has gathered a unique and massive data set of over 1 billion monthly devices, who collectively perform over 2 million sessions per minute on over 2,500 high-traffic news and information websites. Our team of data scientists and machine learning engineers have used this data to unearth the secrets behind online content. Parse.ly serves 300+ enterprise clients with its real-time analytics platform, but the aggregate data exhaust from these integrations provides a front-row seat to what the Internet is looking at today. We'll discuss, through data and visualizations, how consumer attention in the web era really works. We’ll also showcase how we recently applied modern natural language processing techniques to better understand our data set. Throughout, we’ll describe findings related to news trends, social networks, search engines, and device usage patterns. We’ll even discuss some of our predictions about offline consumer behavior, such as which movies would win at the box office based on the web attention those movies received in weeks prior.

Speakers
avatar for Andrew Montalenti

Andrew Montalenti

Founder/CTO, Parse.ly
Co-founder and CTO of Parse.ly, a widely-used real-time web content analytics platform. The product is trusted daily by editors at HuffPost, TIME, TechCrunch, Slate, Quartz, The Wall Street Journal, and 350+ other leading digital companies. Parse.ly helps companies understand the... Read More →

Sponsors
avatar for Capital One

Capital One

Capital One is a diversified bank that offers a broad array of financial products and service to consumers, small businesses and commercial clients. 
avatar for S&P Global

S&P Global

S&P Global Inc. (prior to April 2016 McGraw Hill Financial, Inc., and prior to 2013 McGraw Hill Companies) is an American publicly traded corporation headquartered in New York City. Its primary areas of business are financial information and analytics. It is the parent company of... Read More →


4:00pm

Using Deep Learning to Derive 3D Cities from Satellite Imagery
Limited Capacity full
Adding this to your schedule will put you on the waitlist.

Detection and reconstruction of 3D buildings in urban areas has been a hot topic of research due to its many applications, including 3D population density studies, emergency planning, and building value estimation. Standard approaches to extract building footprint and measure building height rely on either aerial or space borne point cloud data, which in many areas is unavailable. In contrast, high resolution satellite imagery has become more readily available in recent years, and could provide enough information to estimate a building’s height. Recent successes of deep learning for semantic segmentation have shown that convolutional neural networks can be effective tools at extracting 2D building footprints. Using a digital surface model derived from LiDAR data as ground truth, this study goes a step further by employing state of the art deep learning architectures such as U-net to infer both building footprints and estimated building heights in one pass from a single satellite image. This application of open deep learning frameworks can bring the benefits of 3D cities to a larger portion of the world.

Speakers
avatar for Eric Culbertson

Eric Culbertson

Data Scientist, Astaea, Inc.
Eric Culbertson is a Data Scientist at Astraea, Inc. who is fascinated in applying recent advancements in deep learning to increase our understanding of the physical sciences. Before joining Astraea, Eric was an undergraduate physics researcher at both Fermi National Accelerator Laboratory... Read More →

Sponsors
avatar for Capital One

Capital One

Capital One is a diversified bank that offers a broad array of financial products and service to consumers, small businesses and commercial clients. 
avatar for S&P Global

S&P Global

S&P Global Inc. (prior to April 2016 McGraw Hill Financial, Inc., and prior to 2013 McGraw Hill Companies) is an American publicly traded corporation headquartered in New York City. Its primary areas of business are financial information and analytics. It is the parent company of... Read More →


Thursday April 12, 2018 4:00pm - 4:15pm
Violet Crown - Theater 3 200 W Main St, Charlottesville, VA 22902

4:00pm

Understanding the Visual World through Language
Limited Capacity filling up

Building computer vision systems that can recognize objects belonging to a large number of categories is a challenging task for which there has been remarkable progress in recent years. Most of these advances have been fueled by the advent of deep learning and the high availability of annotated datasets. Deep learning has allowed to increase the ability of automated visual recognition systems to perform pattern recognition. However, achieving a more human-like understanding of images is still an open challenge. People use language to describe the visual world, and can reason about objects based on common sense knowledge, domain knowledge, or prior information. I will present a few directions we are pursuing at the vision and language group at the University of Virginia to incorporate these other sources of knowledge into the recognition problem.

Speakers
avatar for Vicente Ordonez

Vicente Ordonez

Assistant Professor, University of Virginia
Vicente Ordonez is Assistant Professor in the Department of Computer Science at the University of Virginia. His research interests lie at the intersection of computer vision, natural language processing and machine learning. He is a recipient of an IBM Faculty Award and a Google Faculty... Read More →

Sponsors
avatar for Capital One

Capital One

Capital One is a diversified bank that offers a broad array of financial products and service to consumers, small businesses and commercial clients. 
avatar for S&P Global

S&P Global

S&P Global Inc. (prior to April 2016 McGraw Hill Financial, Inc., and prior to 2013 McGraw Hill Companies) is an American publicly traded corporation headquartered in New York City. Its primary areas of business are financial information and analytics. It is the parent company of... Read More →


4:15pm

Mapping Deforestation Trends in the Brazilian Amazon Using Machine Learning
Limited Capacity filling up

Monitoring deforestation across the globe is important for understanding macro-scale impacts on carbon sequestration and climate change. Using approximately 2,000 hand-labelled sites scattered across every continent and coarse resolution (500-m) surface reflectance products from the Moderate Resolution Imaging Spectroradiometer (MODIS) satellites, we created a dataset of time windowed features using seven different spectral bands and used these features to train a binary forest classification model to monitor forest change. Our model achieves an estimated accuracy of 95%, a user’s accuracy (recall) of 79%, and a producer’s accuracy (precision) of 95%. Though the model can be applied across the globe, we scored it on a particular area of interest - the State of Mato Grosso in Brazil, which has seen intense deforestation over the past several decades. We mapped and visualized forest change in Mato Grosso over the past 17 years.

Speakers
avatar for Courtney Whalen

Courtney Whalen

Data Scientist, Astraea, Inc.
Courtney Whalen is a data scientist at Astraea, Inc., where she is using satellite imagery and machine learning techniques to answer complex global questions. She has been working as a data scientist for 5 years and has experience developing machine learning models across several... Read More →

Sponsors
avatar for Capital One

Capital One

Capital One is a diversified bank that offers a broad array of financial products and service to consumers, small businesses and commercial clients. 
avatar for S&P Global

S&P Global

S&P Global Inc. (prior to April 2016 McGraw Hill Financial, Inc., and prior to 2013 McGraw Hill Companies) is an American publicly traded corporation headquartered in New York City. Its primary areas of business are financial information and analytics. It is the parent company of... Read More →


Thursday April 12, 2018 4:15pm - 4:30pm
Violet Crown - Theater 3 200 W Main St, Charlottesville, VA 22902

4:20pm

Recommendation Systems for Edge Devices
Limited Capacity seats available

Phone apps, widgets, smart watches, and Alexa; today, machine learning integrated technology is ubiquitous in our homes, cars, devices, and phones. They exist in every domain, from product recommendation to autonomous processes like self-driving vehicles. It is possible for human decisions and choices to both augment and be influenced by machine learning algorithms that are continuously deployed, at scale, to numerous edge devices, allowing users to give and receive real-time (or near real-time) feedback. The process of building and deploying these systems, while on the surface may seem daunting, can be broken down into stages and simplified for a developer, data scientist, or analyst to build on their own. Constructing a fully deployed ML pipeline that communicates openly with the world, can best be demonstrated by something we all enjoy: Movies! A recommendation system utilizing movie data from The Movie Database and user generated data from an iPhone app will be explored both in the mathematics, data curation, and model deployment process. By the end of this talk, a listener will understand the fundamentals of recommendation engine architecture, data required to build one, and how to deploy the model to their iPhone app to collect even more data.

Speakers
avatar for Tyler Hutcherson

Tyler Hutcherson

Data Science Engineer, Metis Machine
Tyler is a data science engineer at local startup, Metis Machine. He acquired a background in physics & mathematical statistics while studying in undergrad at UVa. Recently, his graduate studies culminated with an MS in data science from the UVa Data Science Institute. In his free... Read More →

Sponsors
avatar for Capital One

Capital One

Capital One is a diversified bank that offers a broad array of financial products and service to consumers, small businesses and commercial clients. 
avatar for S&P Global

S&P Global

S&P Global Inc. (prior to April 2016 McGraw Hill Financial, Inc., and prior to 2013 McGraw Hill Companies) is an American publicly traded corporation headquartered in New York City. Its primary areas of business are financial information and analytics. It is the parent company of... Read More →


4:25pm

How to Beat the House - Predicting Football Results with Hyperparameter Optimization
Limited Capacity seats available

The presentation will provide a short overview of machine learning algorithms for classification and techniques to speed up parameter tuning for models and model selection. The dataset used for this has NFL results from 2007 onwards. By selecting the right predictors, algorithm and parameters we are able to predict game results with 70-80% accuracy. This presentation is ideal for beginners who are curious about Data Mining and Machine Learning and would like to start with something fun.

Speakers
avatar for Abhimanyu Roy

Abhimanyu Roy

Student, Data Science Institute at UVA
Abhimanyu Roy is a MS Data Science candidate at the Data Science Institute at the University of Virginia. During his time there he has had the opportunity to work with Capital One's Machine Learning team and with the Center for Leadership Simulation and Gaming at UVA. Prior to this... Read More →

Sponsors
avatar for Capital One

Capital One

Capital One is a diversified bank that offers a broad array of financial products and service to consumers, small businesses and commercial clients. 
avatar for S&P Global

S&P Global

S&P Global Inc. (prior to April 2016 McGraw Hill Financial, Inc., and prior to 2013 McGraw Hill Companies) is an American publicly traded corporation headquartered in New York City. Its primary areas of business are financial information and analytics. It is the parent company of... Read More →


  • Festival (FREE), Included in Summit Badge, or Individually Ticketed Summit

4:30pm

Comparing Baseline Statistical Control to Machine Learning Models in Evaluating Performance in IoT Systems
Limited Capacity seats available

Internet-enabled devices and the Internet of Things (IoT) will continue to become an important component of networked computing systems. These systems leverage “big data” processes that collect, clean, analyze and model large data streams. This project demonstrates techniques and strategies in maintaining performance for IoT systems. Performance is vital to creating reliable data analytics from the edge to the back office. As IoT become more real-time, having reliable performance in the architecture will require near real-time analysis of dependencies for utilization, throughput, resources and access. Establishing baseline performance will become a part of service level agreements (SLAs) of “Big Data” systems. Statistical process control (SPC) and quality improvement in engineering systems that use computing infrastructure have been an industry standard for decades. It has not been, until recently, that such processes were integrated with data processing systems and IP networks. Machine learning (ML) can be used to find anomalies and patterns in the performance of IoT systems by using large datasets to learn from and predict events. The purpose of this project is to compare these two strategies qualitatively and quantitatively while providing guidance for IoT system optimization and monitoring. Business applications for this project include hardware sizing, system health checks, and baseline performance monitoring. The goal is to perform analysis on a series of metrics across multiple layers in an IoT architecture. The Open Systems Interconnection model (OSI model) of IoT will serve as the dichotomy of a baseline performance model. Quantitative and qualitative results will allow businesses to determine scale, performance, accessibility and availability of these systems.

Speakers
avatar for Derek Moore

Derek Moore

Senior Technical Consultant, Independent IT Researcher
Derek Moore is an IT professional with experience in software and services in utilities, telecommunication, manufacturing and energy. A subject matter expert in system analysis and systems engineering, who has successfully implemented enterprise solutions for multinational and national... Read More →

Sponsors
avatar for Capital One

Capital One

Capital One is a diversified bank that offers a broad array of financial products and service to consumers, small businesses and commercial clients. 
avatar for S&P Global

S&P Global

S&P Global Inc. (prior to April 2016 McGraw Hill Financial, Inc., and prior to 2013 McGraw Hill Companies) is an American publicly traded corporation headquartered in New York City. Its primary areas of business are financial information and analytics. It is the parent company of... Read More →


  • Festival (FREE), Included in Summit Badge, or Individually Ticketed Summit

4:35pm

Intro to Wikidata, the Database for Wikipedia and Wiki Projects
Please provide an abstract.For a generation, Wikipedia has been the world's single most popular general information resource, nonprofit media outlet, and source of free and open content. As Wikipedia's almost one billion unique annual users have demanded more and better content, the wiki community has increasingly relied on Wikidata as the wiki database to design every user experience. As with all things "wiki", anyone can edit Wikidata, leading to 8,000 people making multiple contributions to Wikidata's 40 million data items in every month of 2017. Wikidata as a platform offers opportunities for any organization to federate their data into the wiki network as a means to achieve distribution and enjoy crowdsourced data enrichment. This talk showcases projects in art, science, and medicine which anyone could replicate to bring new audiences to their own data.

Speakers
avatar for Lane Rasberry

Lane Rasberry

Wikimedian in Residence, Data Science Institute at UVA
Lane Rasberry is Wikimedian in Residence at the Data Science Institute at the University of Virginia. His interests include consumer rights, open access, crowdsourcing, and coffee.

Sponsors
avatar for Capital One

Capital One

Capital One is a diversified bank that offers a broad array of financial products and service to consumers, small businesses and commercial clients. 
avatar for S&P Global

S&P Global

S&P Global Inc. (prior to April 2016 McGraw Hill Financial, Inc., and prior to 2013 McGraw Hill Companies) is an American publicly traded corporation headquartered in New York City. Its primary areas of business are financial information and analytics. It is the parent company of... Read More →


4:35pm

Location Intelligence at Global Scale: Applying machine learning to maritime, ground transportation, and social media data
Limited Capacity full
Adding this to your schedule will put you on the waitlist.

The proliferation of smart sensors, small satellites, connected cars, and social media have produced a wave of geo-referenced data about human activity. At the same time, the evolution of big data technology and advancements in machine learning have expanded our ability to derive valuable insights about how people interact with each other and their environment. Using satellite collected maritime data from exactEarth, ground transportation data from GPS connected cars and trucks, and exercise data from mobile apps, CCRi CTO Anthony Fox will describe how machine learning techniques can reveal patterns in a broad range of activities: for example, to derive vessel risk scores for insurance analytics, to visualize and assess the impact of major weather events on supply chains, and to discover the best routes for group rides.

Speakers
avatar for Anthony Fox

Anthony Fox

Chief Technology Officer, CCRi
Anthony Fox is the Chief Technology Officer at CCRi. He has over 15 years of experience researching, developing, and productionizing machine learning applications. His recent focus has been on low-latency global scale geo-temporal analytics that combine traditional data fusion with... Read More →

Sponsors
avatar for Capital One

Capital One

Capital One is a diversified bank that offers a broad array of financial products and service to consumers, small businesses and commercial clients. 
avatar for S&P Global

S&P Global

S&P Global Inc. (prior to April 2016 McGraw Hill Financial, Inc., and prior to 2013 McGraw Hill Companies) is an American publicly traded corporation headquartered in New York City. Its primary areas of business are financial information and analytics. It is the parent company of... Read More →


Thursday April 12, 2018 4:35pm - 4:50pm
Violet Crown - Theater 3 200 W Main St, Charlottesville, VA 22902
  • Festival (FREE), Included in Summit Badge, or Individually Ticketed Summit

4:35pm

Imparting Privacy to Face Images: Designing Semi-Adversarial Neural Networks for Multi-Objective Optimization Tasks
Limited Capacity seats available

This talk introduces and discusses a neural network architecture involving a convolutional autoencoder subnetwork capable of perturbing input data such that multiple objectives are optimized. In particular, the proposed architecture is applied to an important challenge in the field of biometrics: protecting privacy in face images while retaining the face recognition utility of biometric identification and verification systems. This novel optimization scheme, semi-adversarial training, is facilitated by a submodule consisting of an auxiliary gender-classifier and an auxiliary face matcher such that an objective function with three terms can be utilized for training: (1) to confound gender attributes of face images, (2) to ensure that the perturbed face images look realistic, and (3), to ensure that biometric face recognition performance is not substantially impacted. Extensive experiments confirm that the resulting model generalizes to multiple, independent face image databases as well as different gender classification and face matching software, showing that the proposed architecture comprises state-of-the-art performance for extending gender privacy to face images. Further, methods for extending this architecture to multiple soft biometric attributes are discussed as well as methods for improving the performance and interpretability of semi-adversarial neural network models.

Speakers
avatar for Sebastian Raschka

Sebastian Raschka

Research, Michigan State University
Sebastian Raschka is a researcher at Michigan State University and Assistant Professor of Statistics at University of Wisconsin-Madison (starting in summer 2018) who develops new methods in computational biology as well as deep learning architectures to solve problems in the field... Read More →

Sponsors
avatar for Capital One

Capital One

Capital One is a diversified bank that offers a broad array of financial products and service to consumers, small businesses and commercial clients. 
avatar for S&P Global

S&P Global

S&P Global Inc. (prior to April 2016 McGraw Hill Financial, Inc., and prior to 2013 McGraw Hill Companies) is an American publicly traded corporation headquartered in New York City. Its primary areas of business are financial information and analytics. It is the parent company of... Read More →


  • Festival (FREE), Included in Summit Badge, or Individually Ticketed Summit

4:40pm

Medical Education Content Categorization with MeSH Terms
Limited Capacity seats available
You need this ticket from Eventbrite to sign up:

Machine Learning Conference Badge

OR

Summit Badge - Advance Badge
Learn how the University of Virginia, School of Medicine VMed program is experimenting with natural language processing (NLP) and artificial intelligence (AI) to automatically categorize educational content, such as exam questions, learning objectives, course materials, images, etc., and attach them to MeSH terms for advanced searching, automatic associations, hierarchical exploration, and recommendation building, in a manner beneficial to both students (for exploring related content that they are having trouble with or want to dive deeper into), and for administrators and faculty (to better audit their content placement with curriculum mapping utilities), opening up course content management to better continuous quality improvement by leveraging algorithms for supervised learning.

Speakers
avatar for Michael Szul

Michael Szul

Principle Software Engineer, University of Virginia, School of Medicine
Michael Szul has spent almost two decades in software engineering in industries as diverse as insurance, energy, travel, and higher education. He has worked in both application and product development, as well as framework creation, code generation, API building, and machine lear... Read More →

Sponsors
avatar for Capital One

Capital One

Capital One is a diversified bank that offers a broad array of financial products and service to consumers, small businesses and commercial clients. 
avatar for S&P Global

S&P Global

S&P Global Inc. (prior to April 2016 McGraw Hill Financial, Inc., and prior to 2013 McGraw Hill Companies) is an American publicly traded corporation headquartered in New York City. Its primary areas of business are financial information and analytics. It is the parent company of... Read More →


4:45pm

4:50pm

Automatically Analyzing Earth Imagery using Deep Learning
Limited Capacity seats available

In the world of earth imagery today there is a huge and embarrassing problem: data is now so cheap and so plentiful that most of it goes completely to waste. Most images of earth captured today are never seen by human eyes, and traditional statistical methods of analysis have an extremely difficult time with the most basic of questions like, "how many cars are in this image," or, "am I looking at a parking lot or an airport?" Only recently has deep learning emerged as a potential answer to this problem. In this talk, we'll touch on how the problem of "too much data" came to be, three examples of how deep learning has been used to automatically analyze massive volumes of earth imagery, and some thoughts on where the future of this industry is going.

Speakers
avatar for Joe Morrison

Joe Morrison

Business Lead, Raster Foundry, Azavea
Joe Morrison leads business development for Raster Foundry, an open source platform for analyzing earth imagery on the web. His job is to help enterprise companies and NGOs to effectively combine earth imagery with analytical techniques, like deep learning, to make more informed decisions... Read More →

Sponsors
avatar for Capital One

Capital One

Capital One is a diversified bank that offers a broad array of financial products and service to consumers, small businesses and commercial clients. 
avatar for S&P Global

S&P Global

S&P Global Inc. (prior to April 2016 McGraw Hill Financial, Inc., and prior to 2013 McGraw Hill Companies) is an American publicly traded corporation headquartered in New York City. Its primary areas of business are financial information and analytics. It is the parent company of... Read More →


Thursday April 12, 2018 4:50pm - 5:05pm
Violet Crown - Theater 3 200 W Main St, Charlottesville, VA 22902

5:00pm

Machine Learning Happy Hour
Limited Capacity full
Adding this to your schedule will put you on the waitlist.
You need this ticket from Eventbrite to sign up: Summit Badge - Advance Badge.
Sponsors
avatar for Capital One

Capital One

Capital One is a diversified bank that offers a broad array of financial products and service to consumers, small businesses and commercial clients. 
avatar for S & P Global

S & P Global

We provide intelligence that is essential for companies, governments and individuals to makedecisions with conviction.


 
Friday, April 13
 

5:00pm

Data and Energy Networking Happy Hour
You need this ticket from Eventbrite to sign up: Summit Badge - Advance Badge.
Grab a drink and swap business cards with other energy industry professionals before heading on to Tom Tom’s Friday Night Block Party.

Sponsors
avatar for Sun Tribe Solar

Sun Tribe Solar

The solar space is dynamic, and keeping in front of advancements can be challenging.  Sun Tribe Solar is a partner with the technical and financial expertise to help your organization take advantage of emerging opportunities to strengthen financial operations while reinforcing your... Read More →


 


Twitter Feed