Schedule

AI Research Week is 5-days of innovation, inspiration and insights featuring notable speakers, panels, workshops, networking and mentorship from leading figures in AI research.

Watch live

September 16 – 20, 2019

Monday

AI Horizons Colloquium Poster Social

AI Horizons Colloquium Poster Session

Time
12:30 PM - 03:00 PM

Location
The MIT Samberg Conference Center

Check out cutting-edge research from the MIT-IBM Watson AI Lab and AI Horizons Network's world-class universities, network with the researchers, and discuss how their work will shape the future of AI. Posters from over 80 collaborative AI research projects will be presented in a social setting with food and drinks provided to promote networking and collaboration building. All AI Research Week participants are welcome to join! Projects cover a variety of topics in AI including fundamental advances in machine learning and reasoning algorithms (deep learning, reinforcement learning, generative adversarial networks, novel NN techniques for program induction, causal structure learning and inference, and many more); AI for healthcare, life sciences, cybersecurity; mapping AI algorithms to quantum and analog architectures; and AI for social good, including ethics and avoiding bias in AI, economics and workforce implications of AI, and AI applied to broad societal challenges. This session will include lightening talks from featured posters and researchers, and awards for the Best Posters. All are welcome! Please register.

AI Mentoring Circles

Time
03:30 PM - 05:00 PM

Location
DR 3-4, 6th floor, MIT Samberg Conference Center

Meet a diversity of leaders in academia and industry at our AI Mentoring Circles, and get guidance on how to successfully build your career in AI.  Each mentoring circle will be a 30-minute topic-focused roundtable that will be held three times during this event.  Our invited mentors will discuss topics in the following themes: Career building in academia vs. industry, professional networking and building collaborations, broadening participation in AI, non-traditional career paths to AI, and balancing work-life demands.

IBM Research Open House 

AI Horizons Colloquium Poster Session

Time
06:00 PM - 09:00 PM

Location
IBM Research Cambridge, 75 Binney Street

Tour the IBM Cambridge Lab, home to the MIT-IBM Watson AI Lab. Check out AI technology demos and presentations, apply for jobs, and take home some swag. Talk to like-minded researchers, students and faculty in a social setting with food and drinks provided.

Tuesday

AI Horizons Colloquium (Invitation only)

AI Horizons Colloquium Poster Session

Time
08:30 AM - 05:30 PM

Location
The MIT Samberg Conference Center 

Join us for our full day marquee event, where leading AI researchers will cover the most compelling issues, questions, and capabilities in AI today. Colloquium presentations and panels bring awareness and understanding of the frontiers of AI, machine learning, deep learning and machine reasoning research; research challenges in making AI and machine learning more robust, fair, and trustworthy; societal challenges that can benefit from AI; and the pressing needs for AI in industries critical to the global economy, such as healthcare, life sciences, finance, and cybersecurity. Presentations will include a fireside chat with Turing Award Winner Professor Yoshua Bengio, and inside views of research projects from our MIT-IBM Watson AI Lab and AI Horizons Network universities by the principal investigators leading those projects.

Wednesday

AI as a Transformative Force for Mental Health Practice and Research

Date
Wednesday, 9/18

Time
08:30 AM - 12:30 PM

Location
The MIT Samberg Conference Center 

Mental illness is prevalent and a major driver of high costs of overall healthcare. One in five Americans suffers from a mental illness that requires care but there is a severe shortage healthcare experts. While AI technology is becoming more prevalent in medical practice, the discipline of mental health has been slower to adopt AI. Mental health practitioners are more hands-on and patient-centered in their clinical practice than most non-psychiatric practitioners, relying more on “softer” skills, like forming relationships with patients and directly observing patient behaviors and emotions. Their clinical data are often in the form of subjective and qualitative patient statements and written notes. AI has great potential to develop digital phenotypes based on behaviors, redefine neuropsychiatric assessment, increase understanding of mental illnesses, and personalize treatment. This workshop will foster a discussion of major challenges in applying AI to mental health and neurological disorders and consider novel strategies to overcome them.  

AI Systems Day

Date
Wednesday, 9/18

Time
09:00 AM - 05:00 PM

Location
The MIT Samberg Conference Center 

This workshop builds on the success of last year’s AI Systems Day workshop targeting the intersection of machine learning and systems. As machine learning techniques rapidly grow in popularity, the design, implementation, and deployment of systems that enable machine learning applications also rapidly grow in importance. Machine learning also feeds back into systems research to provide novel techniques and approaches.   

This workshop has expanded this year to include two parallel tracks and a morning and afternoon session: AI Systems, AI Lifecycle Management, AutoAI Algorithms, and HCI for AutoAI. Topics of interest include both systems for AI and AI for systems, including but not limited to: AutoAI algorithms, HCI of AutoAI and AI, AI lifecycle management, AI platforms, AI programming languages, algorithm toolkits and frameworks, distributed learning, GPU processing, data visualization, AI lifecycle acceleration, AI application composition, automated ML and synthesis, HCI of AI, security and ethics, hardware for AI. 

Enabling Trusted AI

Date
Wednesday, 9/18

Time
09:00 AM - 05:00 PM

Location
The MIT Samberg Conference Center 

AI has made significant advances in the past decade, leading to its usage in various important decision-making scenarios, such as credit scoring, criminal justice, and job recruiting. The trend towards this increased usage in important domains has underscored the need to ensure trust in AI systems.This workshop will focus on this topic of ensuring trust in AI systems.  It will contain a tutorial, invited talks, and a panel. The morning will feature a tutorial on the pillars of Trusted AI and how they connect to the AI lifecycle. It will focus, in detail on two open source toolkits from IBM Research: AI Fairness 360 and AI Explainability 360, that enable researches to explore open research questions and practitioners to improve trust in the AI systems that they build. The afternoon will focus on trustworthy generation of data and models, featuring invited talks and a panel discussion.

MIT-IBM Watson AI Lab Briefing for Prospective Members

Date
Wednesday, 9/18

Time
09:00 AM - 01:00 PM

Location
The MIT Samberg Conference Center 

This invitation-only session will provide interested corporations an in-depth view of the MIT-IBM Watson AI Lab, including the motivations, structure and technical charter for what is the most significant partnership between industry and academic research in the field of Artificial Intelligence.  We’ll review results from the last year, look at new projects that are starting, and even hear about proposed new work that could be highly relevant to banking, telecom, pharma and industrial companies.  The session will also announce the founding members of the Membership Program, where companies can join the MIT-IBM Watson AI Lab by co-investing in exchange for certain benefits.  The session will conclude with a light lunch and networking.

Neuro-Symbolic Computing and Machine Common Sense 

Date
Wednesday, 9/18

Time
09:00 AM - 05:00 PM

Location
The MIT Samberg Conference Center 

Neuro-symbolic methods are relatively new in the DL community and at present, they are largely driven by the findings at the intersection of the fields of child-psychology, generative modelling and neuroscience suggesting symbol manipulation may be at the core of human common sense. This workshop will bring researchers from the aforementioned fields together with the highly diverse set of researchers in the wider fields of representation learning and reasoning in order to advance discussion and promote collaborations across its various subdomains, with a special focus on Neuro-symbolic AI and Neuro-Symbolic Computing and Machine Common Sense.

NASA’s Destination Station and ISS AI Project Pitches!

nasa

Date
Wednesday, 9/18

Time
01:00 PM - 06:00 PM

Location
The MIT Samberg Conference Center 

Partners



The future of AI meets the future of space.

Join us for an exciting opportunity to learn about the world’s only crewed, multinational research laboratory and technology test bed in orbit: the International Space Station (ISS). As part of NASA’s Destination Station, this is a unique opportunity to understand the space-based orbiting laboratory that enables innovative research capable of pushing the boundaries of exploration, and benefitting life on Earth. During this glimpse into the space station we will be joined by representatives from the NASA ISS Program Science Office and ISS U.S. National Laboratory. They will describe the wide range of projects that leverage the space environment, the facilities aboard the orbiting research laboratory, and the vast repository of data collected to date. 

Additionally, you will have the opportunity to meet an astronaut, see and touch artifacts and space food, and hear from researchers and be inspired by ways the future of AI and the future of research in space can intersect. 

 

The event will include a competition and special session in which researchers and students can pitch “ISS meets AI” project ideas of their own to a panel of space experts who will give feedback based on their experience with other flight experiments that have been conducted on the space station.  Awards will be given for the top ideas, and some researchers may be contacted for follow-up discussions.    

More information on the event and how to submit ideas is available at:

Thursday

AI Meets Security Symposium '19 

Date
Thursday, 9/19

Time
09:00 AM - 05:00 PM

Location
MIT Stratton Student Center

As cyber security threats become more sophisticated, stealthy, and devastating, security operations teams struggle to keep up with detecting, managing, and countering cyber attacks -- as well as proactively deploying protective measures. Practitioners are experimenting with leveraging AI technologies in different areas of security operations, including identifying security relevant (mis)behaviors and malware; extracting and consolidating threat intelligence; reasoning over security alerts; recommending countermeasures; and cyber deceptive measures.  At the same time, adversarial attacks on machine learning (ML) systems have become an indisputable threat. Attackers can compromise training of ML models by injecting malicious data into the training set (poisoning attacks), or by crafting adversarial samples that exploit the blind spots of ML models at test time (evasion attacks). This workshop will gain a better understanding of adversarial attacks, and developing more effective defense systems and methods by leveraging AI and ML systems.  This workshop includes a hands on tutorial of the IBM Adversarial Robustness Toolbox (ART) for assessing and hardening AI models. 

KR2ML @ IBM: Knowledge Representation and Reasoning Meets Machine Learning

Date
Thursday, 9/19

Time
09:00 AM - 05:00 PM

Location
The MIT Samberg Conference Center 

Building on the popular KR2ML workshop at AI Research Week 2018, we will gather experts in KRR and machine learning to explore the union of breakthrough methods across the two fields and the outstanding problems that require a level of knowledge and reasoning beyond what is currently available in existing approaches in learning and reasoning. For example, question answering over textual data has made tremendous progress in recent times owing to the advent of neural models such as BERT and ULMFIT. These models are robust to variations in natural language but require large training data and lack the explainability that symbolic models offer. We posit that hybrid methods that utilize the neural models for converting natural language queries into structured representations and symbolic models that use the resultant representations can prove to be useful for question answering systems with wide applications in industry. The workshop will also include an open challenge (released in advance) on translating natural language queries into SPARQL using structured and unstructured data.  

Bridging causal inference, reinforcement learning and transfer learning

Date
Thursday, 9/19

Time
09:00 AM - 03:00 PM

Location
The MIT Samberg Conference Center 

Causal inference is an increasingly popular research direction, focused on discovering causal relations from data and exploiting them to predict the effect of actions/interventions in a system. Recently, there has been exciting new works pointing at connections between causal inference and two other important fields of machine learning: reinforcement learning and transfer learning. Although there seems to be a natural connection between these fields, the different research communities are still separate, a situation complicated by the different terminologies and assumptions. In this workshop we will bring together these different communities in the context of causal inference. We will include speakers from academia and industry that have been pioneering research in these intersections. We will also have a poster session highlighting the work done at IBM and from universities in the Boston Area. 

IBM-WiML Partner Event: New England Women in AI Workshop

Date
Thursday, 9/19

Time
03:00 PM - 06:00 PM

Location
The MIT Samberg Conference Center 

The New England Women in AI Workshop is the flagship event of the IBM Research Cambridge-WiML, partner event series. Quarterly, we hold events with a goal of encouraging women who are students, post-docs or early career researchers in machine learning and AI to build and navigate compelling careers. We do so by offering seminars from thought-leading women researchers and engineers, as well as opportunities for early career researchers to present their own research, to meet and interact with other women in the field, and to find mentors, role models and colleagues. This year, we are pleased to offer talks, a panel, and a networking reception highlighting the work of women in AI. All are welcome to attend. 

Friday

Foundations of Safe Learning

Date
Friday, 9/20

Time
09:00 AM - 1:00 PM

Location
The MIT Samberg Conference Center 

Speakers
Wenchao Li, Bo Li, Aleksander Madry, Stefanie Jegelka, Luca Daniel, Yanzhi Wang 

As deep learning moves from the lab into real-world applications, ensuring the correctness and robustness of deep neural networks becomes a paramount concern. Specifying what it means for a neural network to behave correctly is a challenging problem, especially for classifiers and generative models. Verifying that deep networks meet these specifications is computationally challenging. Recent demonstrations of brittleness in deep learning – including adversarial examples and RL agents that learn pathological control policies – have motivated new computationally tractable approaches toward both specifying and verifying salient properties of neural networks. This workshop will bring together academics, industry researchers, and practitioners to delve into the current state of safe learning. Relevant topics include robustness evaluation/verification, safe reinforcement learning, fairness, robustness against model compression, interpretability, and deep learning with invariance and equivariance constraints. 

Practical Bayesian Methods for Big Data

Date
Friday, 9/20

Time
08:30 AM - 04:30 PM

Location
The MIT Samberg Conference Center 

Speakers
Natesh Pillai, Tamara Broderick, Finale Doshi-Velez, Jan-Willem van de Meent 

Bayesian methods have long benefited from their ability to both coherently represent uncertainty and incorporate prior knowledge, but have traditionally struggled to scale to both large data and large models. Deep learning approaches empirically demonstrate the benefits of learning large over-parameterized models from large data, but struggle with producing well calibrated uncertainties. Research attempting to both scale up Bayesian methods and combine the benefits of either paradigm has recently garnered significant attention. Examples include deep generative models and Bayesian neural networks. This workshop will advance and accelerate research on statistical underpinnings of methods at this intersection, including recent advances in Bayesian approaches for learning neural network based models , deep learning methods for Bayesian modeling, methods for scaling up Bayesian inference to large models and data, and use of classical statistical tools for measuring robustness and reliability of deep learning models.

Designing AI for People

Date
Friday, 9/20

Time
09:00 AM - 04:00 PM

Location
The MIT Samberg Conference Center 

Human-Computer Interaction (HCI) is a multidisciplinary field focusing on the design of computer technology and the interaction between humans and computers. HCI practitioners create consistent interfaces with predictable behaviors: given an input, the system should always produce the same output. AI systems may break this guideline as they create personalized experiences and can learn over time. From decades of HCI research, we know that inconsistent and unpredictable behaviors can confuse users, erode their confidence, and lead to abandonment of a technology. In this hands-on tutorial, participants will design an AI-infused product/service using design thinking methods. They will learn to identify and build empathy with their users, articulate and prioritize their needs, design solutions, and evaluate how the incorporation of AI may impact users’ trust in and use of the product/service. They will walk away with a strong understanding of the importance of including end-users in the design process, practical exercises for eliciting user needs, and ways of understanding how to design and evaluate AI-infused systems. 

Question Answering and Semantic Parsing (QASP) Workshop

Date
Friday, 9/20

Time
09:00 AM - 04:00 PM

Location
The MIT Samberg Conference Center 

Machine Reading and Question Answering (MRQA) is an important research topic for evaluating how well AI systems understand natural language, and also  critical for industry applications such as search engines and speech and dialogue systems. In a typical MRQA setup, a system must answer a question by reading one or more context documents. Successful MRQA systems must understand a wide range of natural language situations, and a wide variety of question and document types. Meanwhile, semantic parsers map sentences to formal representations of their underlying meaning. Recently, algorithms have been developed to learn to recover increasingly expressive representations with ever weaker forms of supervision. This workshop aims to achieve two goals. First, to bring together researchers (university and industry) in the field to discuss the state of the art and opportunities for future research. Second, to create a stage for presenting the variety of current approaches, thereby providing a unique opportunity for new entrants to the field. 

Speakers

AI Research Week features keynotes, panels, and workshops lead by a range of distinguished speakers.

Questions about AI Research Week?

Have a question about any of the events at AI Research Week, hotel accommodation, or need help registering?

Learn more about IBM Research AI

AI Research Week replays

Relive the innovation, inspiration, and insights of AI Research Week 2018

MIT-IBM Watson AI Lab

A collaborative industrial-academic laboratory focused on advancing fundamental AI research.

Try our tech

At IBM Research, we create innovative tools and resources to help unleash the power of AI. See for yourself. Take our tech for a spin.

AI Horizons Colloquium schedule

8:00 AM

Registration / Check-in

8:30 AM

Welcome Address

8:40 AM

IBM Research and the Future of Computing

Dario Gil
Director, IBM Research, IBM Chair, MIT-IBM Watson AI Lab

9:05 AM

Keynote: Beyond IID: Meta-learning Disentangled Causal Variables and How the World Works

Yoshua Bengio
A.M. Turing Award Full Professor, Department of Computer Science and Operations Research, Université de Montréal Canada Research Chair in Statistical Learning Algorithms, Founder and Scientific Director of Mila, Scientific Director of IVADO, CIFAR Fellow and Program Director

10:00 AM

How Can We Trust AI?

Saska Mojsilovic
Head of AI Foundations, IBM Research Co-Director of IBM Science for Social Good, IBM Fellow

10:30 PM

Break

10:45 AM

Clinically Useful Machine Learning: How Close Are We?

Collin Stultz
Professor of Electrical Engineering and Computer Science Institute for Medical Engineering and Science, MIT Faculty at Harvard-MIT Division of Health Sciences and Technology Cardiologist, Massachusetts General Hospital (MGH)

11:15 AM

From AI Research to Industries

David Cox
IBM Director, MIT-IBM Watson AI Lab

Aude Oliva
MIT Executive Director, MIT-IBM Watson AI Lab MIT Quest for Intelligence

11:45 AM

Mastering Language

Roger Levy
Associate Professor in the Department of Brain and Cognitive Sciences at MIT, Director of the Computational Psycholinguistics Laboratory

12:15 PM

Lunch and Networking

1:15 PM

Self-supervised Learning

Antonio Torralba
Professor of Electrical Engineering and Computer Science at MIT, MIT Director of the MIT-IBM Watson AI Lab, Inaugural Director of the MIT Quest for Intelligence

1:45 PM

Spotlight talks from AI Horizons Network and MIT-IBM Watson AI Lab Collaborative Research Projects

Michael Carbin
Assistant Professor of Electrical Engineering and Computer Science at MIT

Chris Sims
Assistant Professor, Department of Cognitive Science, Rensselaer Polytechnic Institute

Julie Shah
Associate Professor, Department of Aeronautics and Astronautics at MIT Associate Professor, Computer Science and Artificial Intelligence Lab at MIT

Alexander Schwing
Assistant Professor in the Department of Electrical and Computer Engineering at the University of Illinois in Urbana-Champaign

Hendrik Strobelt
Research Staff Member, IBM Research

03:30 PM

Break

3:45 PM

Panel: Towards more Human-like learning in AI

Chair
Josh Tenenbaum
Professor, Dept of Brain and Cognitive Sciences, MIT

Panelists
Laura Shulz
Professor of Cognitive Science in the Brain and Cognitive Sciences department at MIT, Primary Investigator of the MIT Early Childhood Cognition Lab

Leslie Kaebling
Panasonic Professor of Computer Science and Engineering

Pulkit Agrawal
Assistant Professor of Electrical Engineering and Computer Science at MIT

4:30 PM

Perspectives on the Future of Computing for AI

Andrew McCallum
Distinguished Professor of Computer Science, University of Massachusetts at Amherst

Song Han
Assistant Professor of Electrical Engineering and Computer Science at MIT

Jerry Chow
IBM, Manager, Experimental Quantum Computing

Vivienne Sze
Associate Professor of Electrical Engineering and Computer Science, MIT

 

5:15 PM

Wrap up and Adjourn