{"id":712540,"date":"2021-03-12T13:34:28","date_gmt":"2021-03-12T21:34:28","guid":{"rendered":"https:\/\/www.microsoft.com\/en-us\/research\/?post_type=msr-event&p=712540"},"modified":"2025-08-06T11:51:42","modified_gmt":"2025-08-06T18:51:42","slug":"msr-mila-workshop-2021","status":"publish","type":"msr-event","link":"https:\/\/www.microsoft.com\/en-us\/research\/event\/msr-mila-workshop-2021\/","title":{"rendered":"MSR & Mila Workshop 2021"},"content":{"rendered":"\n\n

This is an invite-only event.<\/strong><\/p>\n

Sponsor:<\/strong> Microsoft Research Lab \u2013 Montreal<\/a>Opens in a new tab<\/span><\/p>\n

You\u2019re invited to join us for the 2021 MSR & Mila Research Workshop held virtually on Microsoft Teams. The workshop will feature presentations and forward-thinking breakout sessions by Mila professors, students and researchers at Microsoft.<\/p>\n

The workshop is an opportunity for Mila professors, students and MSR researchers to share and present what they\u2019re currently working on with the goal of facilitating and building on existing research collaborations between Mila & MSR.<\/p>\n

Organizing Committee<\/h3>\n

Brittney Muller, Community Engagement Program Manager, Microsoft Research Outreach
\nGeoff Gordon, Partner Researcher, Microsoft Research Montreal
\nJessica Mastronardi, Senior Manager Community Engagement & Partnerships, Microsoft Research Outreach
\nSusan Dumais, Technical Fellow and Lab Director, Microsoft Research New England, New York City and Montreal
\nYoshua Bengio, Founder and Scientific Director of Mila<\/p>\n

Microsoft\u2019s Event Code of Conduct<\/h3>\n

Microsoft\u2019s mission is to empower every person and every organization on the planet to achieve more. This includes virtual events Microsoft hosts and participates in, where we seek to create a respectful, friendly, and inclusive experience for all participants. As such, we do not tolerate harassing or disrespectful behavior, messages, images, or interactions by any event participant, in any form, at any aspect of the program including business and social activities, regardless of location.<\/p>\n

We do not tolerate any behavior that is degrading to any gender, race, sexual orientation or disability, or any behavior that would violate Microsoft\u2019s Anti-Harassment and Anti-Discrimination Policy, Equal Employment Opportunity Policy, or Standards of Business Conduct (opens in new tab)<\/span><\/a>. In short, the entire experience must meet our culture standards. We encourage everyone to assist in creating a welcoming and safe environment. Please report (opens in new tab)<\/span><\/a> any concerns, harassing behavior, or suspicious or disruptive activity. Microsoft reserves the right to ask attendees to leave at any time at its sole discretion.<\/p>\n

<\/div>\n
\n\t\n\t\tReport a concern\t<\/a>\n\n\t<\/div>\n

Opens in a new tab<\/span><\/p>\n

Tuesday, April 13<\/h2>\n\n\n\n\n\n\n\n\n\n\n\n
Time (EST)<\/strong><\/td>\nSession<\/strong><\/td>\n<\/td>\n<\/tr>\n
12:30 PM-12:35 PM<\/td>\nWelcome<\/strong>
\nJessica Mastronardi, Senior Manager Community Engagement & Partnerships, Microsoft Research Outreach
\nSusan Dumais, Technical Fellow and Lab Director, Microsoft Research Montreal, New England, and NYC<\/td>\n		<\/td>\n<\/tr>\n
12:35 PM-1:05 PM<\/td>\nInteractive Representation Learning<\/strong>
\nDevon Hjelm, Principal Researcher, Microsoft Research Montreal<\/td>\n		<\/td>\n<\/tr>\n
1:05 PM-1:35 PM<\/td>\nChatbots: There and Back Again<\/strong>
\nMahmoud Adada, Principal Engineer Manager, Turing Montreal<\/td>\n		<\/td>\n<\/tr>\n
1:35 PM-2:05 PM<\/td>\nMachines with Common Sense? Modelling Plausibility in Text<\/strong>
\nJackie Cheung, Assistant Professor, McGill University, Mila and Consulting Researcher, Microsoft Research Montreal<\/td>\n		<\/td>\n<\/tr>\n
2:05 PM-2:15 PM<\/td>\n10 Minute Break<\/td>\n<\/td>\n<\/tr>\n
2:15 PM-3:25 PM<\/td>\nBreakout Session A: Generalization Failures in Deep Nets<\/strong>
\nAlessandro Sordoni, Principal Researcher, Microsoft Research
\nSara Hooker, Google Brain
\nGauthier Gidel, Universit\u00e9 de Montr\u00e9al and Mila
\nRemi Tachet des Combes, Microsoft Research
\nAaron Courville, Universit\u00e9 de Montr\u00e9al and Mila<\/td>\n		Breakout Session B: FATE concerns in NLP<\/strong>
\nAlexandra Olteanu, Principal Researcher, Microsoft Research
\nSu Lin Blodgett, Post Doc Researcher, Microsoft Research
\nHal Daume III, Microsoft Research
\nSolon Barocas, Microsoft Research
\nJackie Cheung, Mcgill University and Mila
\nAmandalynne Paullada, University of Washington<\/td>\n<\/tr>\n
3:25 PM-3:55 PM<\/td>\nDeep Learning of Abstract Mechanisms, Abstract Entities and Abstract Actions<\/strong>
\nProfessor Yoshua Bengio, Founder and Scientific Director of Mila<\/td>\n		<\/td>\n<\/tr>\n
3:55 PM-4:00 PM<\/td>\nClosing Discussion and Thank You Remarks<\/strong>
\nProfessor Yoshua Bengio, Founder and Scientific Director of Mila
\nPeter Lee, CVP, Research and Incubations, Microsoft Research<\/td>\n		<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n
<\/div>\n

Wednesday, April 14<\/h2>\n\n\n\n\n\n\n\n\n\n\n\n
Time (EST)<\/strong><\/td>\nSession<\/strong><\/td>\n<\/td>\n<\/tr>\n
12:30 PM-12:35 PM<\/td>\nWelcome<\/strong>
\nJessica Mastronardi, Senior Manager Community Engagement & Partnerships, Microsoft Research Outreach
\nSusan Dumais, Technical Fellow and Lab Director, Microsoft Research Montreal, New England, and NYC<\/td>\n		<\/td>\n<\/tr>\n
12:35 PM-1:05 PM<\/td>\nTowards Hyperparameter-Free Optimization for Modern Machine Learning<\/strong>
\nSimon Lacoste-Julien, Associate Professor, Universit\u00e9 de Montr\u00e9al, Mila<\/td>\n		<\/td>\n<\/tr>\n
1:05 PM-1:35 PM<\/td>\nSuccessor Feature Sets for Planning and Imitation<\/strong>
\nGeoff Gordon, Partner Researcher, Microsoft Research Montreal<\/td>\n		<\/td>\n<\/tr>\n
1:35 PM-2:05 PM<\/td>\nDialogue as Collaborative Construction of Dataflow<\/strong>
\nJason Eisner, Partner Principal Researcher, Microsoft Semantic Machines<\/td>\n		<\/td>\n<\/tr>\n
2:05 PM-2:15 PM<\/td>\n10 Minute Break<\/td>\n<\/td>\n<\/tr>\n
2:15 PM-3:25 PM<\/td>\nBreakout Session A: Memory in RL<\/strong>
\nIda Momennejad, Senior Researcher, Microsoft Research
\nBlake Richards, Assistant Professor, McGill University, Mila
\nGuillaume Lajoie, Assistant Professor, Universit\u00e9 de Montr\u00e9al, Mila<\/td>\n		Breakout Session B: Observe, Remember, and Assist<\/strong>
\nAndrew McNamara, Director of Engineering, Turing
\nGeoff Gordon, Microsoft Research
\nGerald Haslhofer, Microsoft (MMX)
\nKaheer Suleman, Microsoft Research
\nTim Franklin, Microsoft (MMX)
\nTianyu Li, McGill University<\/td>\n<\/tr>\n
3:25 PM-3:55 PM<\/td>\nTowards Broad and Robust AI<\/strong>
\nIrina Rish, Associate Professor, University of Montreal, Mila<\/td>\n		<\/td>\n<\/tr>\n
3:55 PM-4:00 PM<\/td>\nClosing Discussion and Thank You Remarks<\/strong>
\nProfessor Yoshua Bengio, Founder and Scientific Director of Mila
\nPeter Lee, CVP, Research and Incubations, Microsoft Research<\/td>\n		<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

Opens in a new tab<\/span><\/p>\n

Tuesday, April 13 | 2:15 PM – 3:25 PM EST<\/h2>\n

One the day of the workshop, select one of the following breakout sessions to attend.<\/p>\n

Breakout Session A: Generalization Failures in Deep Nets<\/strong><\/h3>\n

Facilitator:
\nAlessandro Sordoni, Microsoft Research Montreal
\nRemi Tachet des Combes, Microsoft Research Montreal<\/p>\n

Panelists:
\nSara Hooker, Google Brain
\nGauthier Gidel, Universit\u00e9 de Montr\u00e9al and Mila
\nAaron Courville, Universit\u00e9 de Montr\u00e9al and Mila<\/p>\n

Abstract:
\nAlthough large (overparametrized) deep neural networks do wonders, level-5 autonomous driving appears still out of reach. In this short and speculative session, we won’t brainstorm about autonomous driving (yet), but instead will question how deep nets behave in the long tail, on adversarial examples, with spurious correlations and more generally out-of-distribution. We will hear about some generalization failures of deep nets as well as recent approaches that have been explored to fix them. This breakout is intended to be for both Neural Nets lovers and skeptics<\/p>\n

This session will start with a panel discussion, followed by a live Q&A with participants.<\/p>\n

Breakout Session B: FATE Concerns in NLP<\/strong><\/h3>\n

Facilitators:
\nAlexandra Olteanu, Microsoft Research Montreal
\nSu Lin Blodgett, Microsoft Research Montreal<\/p>\n

Panelists:
\nAmandalynne Paullada, University of Washington
\nHal Daume III, Microsoft Research New York City
\nJackie Cheung, McGill University and Mila
\nSolon Barocas, Microsoft Research New York City<\/p>\n

Abstract:
\nIn this breakout session we will discuss existing concerns related to the fairness, accountability, transparency and ethics (FATE) of language technologies (and the data they ingest or generate), as these technologies might surface, replicate, exacerbate or even cause a range of computational harms (from surfacing offensive speech or reinforcing stereotypes, to even more subtle issues like nudging users towards undesirable patterns of behavior or triggering memories of traumatic events). How can we reliably measure fairness related harms? Whose data is included when training the model? How do we better foresee potential harms? How do we engage stakeholders? And many other critical questions.<\/p>\n

This session will start with a panel discussion, followed by a live Q&A with participants.<\/p>\n

Wednesday, April 14 | 2:15 PM – 3:25 PM EST<\/h2>\n

One the day of the workshop, select one of the following breakout sessions to attend.<\/p>\n

Breakout Session A: Memory in RL<\/strong><\/h3>\n

Facilitators:
\nBlake Richards, McGill University and Mila
\nGuillaume LaJoie, Universit\u00e9 de Montr\u00e9al and Mila
\nIda Momennejad, Microsoft Research NYC<\/p>\n

Abstract:
\nThe early development of reinforcement learning (RL) was directly inspired by ideas from neuroscience and psychology, but over time, RL produced theories and techniques that bore less and less resemblance to mechanisms in the brain. However, there has been growing interest in the RL community to re-examine the potential for insights about memory in the brain to inform RL. In particular, the brain’s declarative memory systems (episodic and semantic) may provide guidance as to how to build RL systems that can avoid catastrophic forgetting, engage in long-distance credit assignment, and generalize to novel situations. In this workshop we will discuss the work to-date on memory systems for RL, some new insights into memory from neuroscience, and potential future directions in this area.<\/p>\n

This session will start with a topic overview, followed by a live discussion and Q&A with participants.<\/p>\n

Breakout Session B: Observe, Remember, and Assist<\/strong><\/h3>\n

Facilitator:
\nAndrew McNamara, Turing Montreal<\/p>\n

Speakers:
\nTalk 1: Observe Remember Automation: AI-Assisted Productivity
\nTim Franklin, Microsoft<\/p>\n

Talk 2: Automating Style Transfer in Office Documents
\nTianyu Li, McGill University<\/p>\n

Panelists:
\nGeoff Gordon, Microsoft Research
\nGerald Haslhofer, Microsoft
\nKaheer Suleman, Microsoft Research<\/p>\n

Abstract:
\nObserve, Remember, and Automate is a formulation of remote process automation whereby we seek to empower users to achieve more. Our goal is to automate as much as possible for users to limit their time on mundane and repetitive tasks. We want to help users optimize their time by making recommendations and predictions based on real time context and actions taking into account everything we’ve observed and remembered from past interactions. This effort requires the combination of various research areas including language modeling, representation learning, computer vision, reinforcement learning, graph representations, and much more.<\/p>\n

This session will start with a topic overview, followed by a live discussion and Q&A with participants.Opens in a new tab<\/span><\/p>\n","protected":false},"excerpt":{"rendered":"

This is an invite-only event. Sponsor: Microsoft Research Lab \u2013 MontrealOpens in a new tab You\u2019re invited to join us for the 2021 MSR & Mila Research Workshop held virtually on Microsoft Teams. The workshop will feature presentations and forward-thinking breakout sessions by Mila professors, students and researchers at Microsoft. The workshop is an opportunity […]<\/p>\n","protected":false},"featured_media":718087,"template":"","meta":{"msr-url-field":"","msr-podcast-episode":"","msrModifiedDate":"","msrModifiedDateEnabled":false,"ep_exclude_from_search":false,"_classifai_error":"","msr_startdate":"","msr_enddate":"","msr_location":"Virtual (Microsoft Teams)","msr_expirationdate":"","msr_event_recording_link":"","msr_event_link":"","msr_event_link_redirect":false,"msr_event_time":"","msr_hide_region":false,"msr_private_event":true,"msr_hide_image_in_river":0,"footnotes":""},"research-area":[13556],"msr-region":[197900],"msr-event-type":[197944,210063],"msr-video-type":[],"msr-locale":[268875],"msr-program-audience":[],"msr-post-option":[],"msr-impact-theme":[],"class_list":["post-712540","msr-event","type-msr-event","status-publish","has-post-thumbnail","hentry","msr-research-area-artificial-intelligence","msr-region-north-america","msr-event-type-hosted-by-microsoft","msr-event-type-workshop","msr-locale-en_us"],"msr_about":"\n\n

This is an invite-only event.<\/strong><\/p>\n

Sponsor:<\/strong> Microsoft Research Lab \u2013 Montreal<\/a>Opens in a new tab<\/span><\/p>\n

You\u2019re invited to join us for the 2021 MSR & Mila Research Workshop held virtually on Microsoft Teams. The workshop will feature presentations and forward-thinking breakout sessions by Mila professors, students and researchers at Microsoft.<\/p>\n

The workshop is an opportunity for Mila professors, students and MSR researchers to share and present what they\u2019re currently working on with the goal of facilitating and building on existing research collaborations between Mila & MSR.<\/p>\n

Organizing Committee<\/h3>\n

Brittney Muller, Community Engagement Program Manager, Microsoft Research Outreach
\nGeoff Gordon, Partner Researcher, Microsoft Research Montreal
\nJessica Mastronardi, Senior Manager Community Engagement & Partnerships, Microsoft Research Outreach
\nSusan Dumais, Technical Fellow and Lab Director, Microsoft Research New England, New York City and Montreal
\nYoshua Bengio, Founder and Scientific Director of Mila<\/p>\n

Microsoft\u2019s Event Code of Conduct<\/h3>\n

Microsoft\u2019s mission is to empower every person and every organization on the planet to achieve more. This includes virtual events Microsoft hosts and participates in, where we seek to create a respectful, friendly, and inclusive experience for all participants. As such, we do not tolerate harassing or disrespectful behavior, messages, images, or interactions by any event participant, in any form, at any aspect of the program including business and social activities, regardless of location.<\/p>\n

We do not tolerate any behavior that is degrading to any gender, race, sexual orientation or disability, or any behavior that would violate Microsoft\u2019s Anti-Harassment and Anti-Discrimination Policy, Equal Employment Opportunity Policy, or Standards of Business Conduct (opens in new tab)<\/span><\/a>. In short, the entire experience must meet our culture standards. We encourage everyone to assist in creating a welcoming and safe environment. Please report (opens in new tab)<\/span><\/a> any concerns, harassing behavior, or suspicious or disruptive activity. Microsoft reserves the right to ask attendees to leave at any time at its sole discretion.<\/p>\n

<\/div>\n
\n\t\n\t\tReport a concern\t<\/a>\n\n\t<\/div>\n

Opens in a new tab<\/span><\/p>\n

Tuesday, April 13<\/h2>\n\n\n\n\n\n\n\n\n\n\n\n
Time (EST)<\/strong><\/td>\nSession<\/strong><\/td>\n<\/td>\n<\/tr>\n
12:30 PM-12:35 PM<\/td>\nWelcome<\/strong>
\nJessica Mastronardi, Senior Manager Community Engagement & Partnerships, Microsoft Research Outreach
\nSusan Dumais, Technical Fellow and Lab Director, Microsoft Research Montreal, New England, and NYC<\/td>\n		<\/td>\n<\/tr>\n
12:35 PM-1:05 PM<\/td>\nInteractive Representation Learning<\/strong>
\nDevon Hjelm, Principal Researcher, Microsoft Research Montreal<\/td>\n		<\/td>\n<\/tr>\n
1:05 PM-1:35 PM<\/td>\nChatbots: There and Back Again<\/strong>
\nMahmoud Adada, Principal Engineer Manager, Turing Montreal<\/td>\n		<\/td>\n<\/tr>\n
1:35 PM-2:05 PM<\/td>\nMachines with Common Sense? Modelling Plausibility in Text<\/strong>
\nJackie Cheung, Assistant Professor, McGill University, Mila and Consulting Researcher, Microsoft Research Montreal<\/td>\n		<\/td>\n<\/tr>\n
2:05 PM-2:15 PM<\/td>\n10 Minute Break<\/td>\n<\/td>\n<\/tr>\n
2:15 PM-3:25 PM<\/td>\nBreakout Session A: Generalization Failures in Deep Nets<\/strong>
\nAlessandro Sordoni, Principal Researcher, Microsoft Research
\nSara Hooker, Google Brain
\nGauthier Gidel, Universit\u00e9 de Montr\u00e9al and Mila
\nRemi Tachet des Combes, Microsoft Research
\nAaron Courville, Universit\u00e9 de Montr\u00e9al and Mila<\/td>\n		Breakout Session B: FATE concerns in NLP<\/strong>
\nAlexandra Olteanu, Principal Researcher, Microsoft Research
\nSu Lin Blodgett, Post Doc Researcher, Microsoft Research
\nHal Daume III, Microsoft Research
\nSolon Barocas, Microsoft Research
\nJackie Cheung, Mcgill University and Mila
\nAmandalynne Paullada, University of Washington<\/td>\n<\/tr>\n
3:25 PM-3:55 PM<\/td>\nDeep Learning of Abstract Mechanisms, Abstract Entities and Abstract Actions<\/strong>
\nProfessor Yoshua Bengio, Founder and Scientific Director of Mila<\/td>\n		<\/td>\n<\/tr>\n
3:55 PM-4:00 PM<\/td>\nClosing Discussion and Thank You Remarks<\/strong>
\nProfessor Yoshua Bengio, Founder and Scientific Director of Mila
\nPeter Lee, CVP, Research and Incubations, Microsoft Research<\/td>\n		<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n
<\/div>\n

Wednesday, April 14<\/h2>\n\n\n\n\n\n\n\n\n\n\n\n
Time (EST)<\/strong><\/td>\nSession<\/strong><\/td>\n<\/td>\n<\/tr>\n
12:30 PM-12:35 PM<\/td>\nWelcome<\/strong>
\nJessica Mastronardi, Senior Manager Community Engagement & Partnerships, Microsoft Research Outreach
\nSusan Dumais, Technical Fellow and Lab Director, Microsoft Research Montreal, New England, and NYC<\/td>\n		<\/td>\n<\/tr>\n
12:35 PM-1:05 PM<\/td>\nTowards Hyperparameter-Free Optimization for Modern Machine Learning<\/strong>
\nSimon Lacoste-Julien, Associate Professor, Universit\u00e9 de Montr\u00e9al, Mila<\/td>\n		<\/td>\n<\/tr>\n
1:05 PM-1:35 PM<\/td>\nSuccessor Feature Sets for Planning and Imitation<\/strong>
\nGeoff Gordon, Partner Researcher, Microsoft Research Montreal<\/td>\n		<\/td>\n<\/tr>\n
1:35 PM-2:05 PM<\/td>\nDialogue as Collaborative Construction of Dataflow<\/strong>
\nJason Eisner, Partner Principal Researcher, Microsoft Semantic Machines<\/td>\n		<\/td>\n<\/tr>\n
2:05 PM-2:15 PM<\/td>\n10 Minute Break<\/td>\n<\/td>\n<\/tr>\n
2:15 PM-3:25 PM<\/td>\nBreakout Session A: Memory in RL<\/strong>
\nIda Momennejad, Senior Researcher, Microsoft Research
\nBlake Richards, Assistant Professor, McGill University, Mila
\nGuillaume Lajoie, Assistant Professor, Universit\u00e9 de Montr\u00e9al, Mila<\/td>\n		Breakout Session B: Observe, Remember, and Assist<\/strong>
\nAndrew McNamara, Director of Engineering, Turing
\nGeoff Gordon, Microsoft Research
\nGerald Haslhofer, Microsoft (MMX)
\nKaheer Suleman, Microsoft Research
\nTim Franklin, Microsoft (MMX)
\nTianyu Li, McGill University<\/td>\n<\/tr>\n
3:25 PM-3:55 PM<\/td>\nTowards Broad and Robust AI<\/strong>
\nIrina Rish, Associate Professor, University of Montreal, Mila<\/td>\n		<\/td>\n<\/tr>\n
3:55 PM-4:00 PM<\/td>\nClosing Discussion and Thank You Remarks<\/strong>
\nProfessor Yoshua Bengio, Founder and Scientific Director of Mila
\nPeter Lee, CVP, Research and Incubations, Microsoft Research<\/td>\n		<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

Opens in a new tab<\/span><\/p>\n

Tuesday, April 13 | 2:15 PM – 3:25 PM EST<\/h2>\n

One the day of the workshop, select one of the following breakout sessions to attend.<\/p>\n

Breakout Session A: Generalization Failures in Deep Nets<\/strong><\/h3>\n

Facilitator:
\nAlessandro Sordoni, Microsoft Research Montreal
\nRemi Tachet des Combes, Microsoft Research Montreal<\/p>\n

Panelists:
\nSara Hooker, Google Brain
\nGauthier Gidel, Universit\u00e9 de Montr\u00e9al and Mila
\nAaron Courville, Universit\u00e9 de Montr\u00e9al and Mila<\/p>\n

Abstract:
\nAlthough large (overparametrized) deep neural networks do wonders, level-5 autonomous driving appears still out of reach. In this short and speculative session, we won’t brainstorm about autonomous driving (yet), but instead will question how deep nets behave in the long tail, on adversarial examples, with spurious correlations and more generally out-of-distribution. We will hear about some generalization failures of deep nets as well as recent approaches that have been explored to fix them. This breakout is intended to be for both Neural Nets lovers and skeptics<\/p>\n

This session will start with a panel discussion, followed by a live Q&A with participants.<\/p>\n

Breakout Session B: FATE Concerns in NLP<\/strong><\/h3>\n

Facilitators:
\nAlexandra Olteanu, Microsoft Research Montreal
\nSu Lin Blodgett, Microsoft Research Montreal<\/p>\n

Panelists:
\nAmandalynne Paullada, University of Washington
\nHal Daume III, Microsoft Research New York City
\nJackie Cheung, McGill University and Mila
\nSolon Barocas, Microsoft Research New York City<\/p>\n

Abstract:
\nIn this breakout session we will discuss existing concerns related to the fairness, accountability, transparency and ethics (FATE) of language technologies (and the data they ingest or generate), as these technologies might surface, replicate, exacerbate or even cause a range of computational harms (from surfacing offensive speech or reinforcing stereotypes, to even more subtle issues like nudging users towards undesirable patterns of behavior or triggering memories of traumatic events). How can we reliably measure fairness related harms? Whose data is included when training the model? How do we better foresee potential harms? How do we engage stakeholders? And many other critical questions.<\/p>\n

This session will start with a panel discussion, followed by a live Q&A with participants.<\/p>\n

Wednesday, April 14 | 2:15 PM – 3:25 PM EST<\/h2>\n

One the day of the workshop, select one of the following breakout sessions to attend.<\/p>\n

Breakout Session A: Memory in RL<\/strong><\/h3>\n

Facilitators:
\nBlake Richards, McGill University and Mila
\nGuillaume LaJoie, Universit\u00e9 de Montr\u00e9al and Mila
\nIda Momennejad, Microsoft Research NYC<\/p>\n

Abstract:
\nThe early development of reinforcement learning (RL) was directly inspired by ideas from neuroscience and psychology, but over time, RL produced theories and techniques that bore less and less resemblance to mechanisms in the brain. However, there has been growing interest in the RL community to re-examine the potential for insights about memory in the brain to inform RL. In particular, the brain’s declarative memory systems (episodic and semantic) may provide guidance as to how to build RL systems that can avoid catastrophic forgetting, engage in long-distance credit assignment, and generalize to novel situations. In this workshop we will discuss the work to-date on memory systems for RL, some new insights into memory from neuroscience, and potential future directions in this area.<\/p>\n

This session will start with a topic overview, followed by a live discussion and Q&A with participants.<\/p>\n

Breakout Session B: Observe, Remember, and Assist<\/strong><\/h3>\n

Facilitator:
\nAndrew McNamara, Turing Montreal<\/p>\n

Speakers:
\nTalk 1: Observe Remember Automation: AI-Assisted Productivity
\nTim Franklin, Microsoft<\/p>\n

Talk 2: Automating Style Transfer in Office Documents
\nTianyu Li, McGill University<\/p>\n

Panelists:
\nGeoff Gordon, Microsoft Research
\nGerald Haslhofer, Microsoft
\nKaheer Suleman, Microsoft Research<\/p>\n

Abstract:
\nObserve, Remember, and Automate is a formulation of remote process automation whereby we seek to empower users to achieve more. Our goal is to automate as much as possible for users to limit their time on mundane and repetitive tasks. We want to help users optimize their time by making recommendations and predictions based on real time context and actions taking into account everything we’ve observed and remembered from past interactions. This effort requires the combination of various research areas including language modeling, representation learning, computer vision, reinforcement learning, graph representations, and much more.<\/p>\n

This session will start with a topic overview, followed by a live discussion and Q&A with participants.Opens in a new tab<\/span><\/p>\n","tab-content":[{"id":0,"name":"About","content":"You\u2019re invited to join us for the 2021 MSR & Mila Research Workshop held virtually on Microsoft Teams. The workshop will feature presentations and forward-thinking breakout sessions by Mila professors, students and researchers at Microsoft.\r\n\r\nThe workshop is an opportunity for Mila professors, students and MSR researchers to share and present what they\u2019re currently working on with the goal of facilitating and building on existing research collaborations between Mila & MSR.\r\n

Organizing Committee<\/h3>\r\nBrittney Muller, Community Engagement Program Manager, Microsoft Research Outreach\r\nGeoff Gordon, Partner Researcher, Microsoft Research Montreal\r\nJessica Mastronardi, Senior Manager Community Engagement & Partnerships, Microsoft Research Outreach\r\nSusan Dumais, Technical Fellow and Lab Director, Microsoft Research New England, New York City and Montreal\r\nYoshua Bengio, Founder and Scientific Director of Mila\r\n

Microsoft\u2019s Event Code of Conduct<\/h3>\r\nMicrosoft\u2019s mission is to empower every person and every organization on the planet to achieve more. This includes virtual events Microsoft hosts and participates in, where we seek to create a respectful, friendly, and inclusive experience for all participants. As such, we do not tolerate harassing or disrespectful behavior, messages, images, or interactions by any event participant, in any form, at any aspect of the program including business and social activities, regardless of location.\r\n\r\nWe do not tolerate any behavior that is degrading to any gender, race, sexual orientation or disability, or any behavior that would violate Microsoft\u2019s Anti-Harassment and Anti-Discrimination Policy, Equal Employment Opportunity Policy, or Standards of Business Conduct<\/a>. In short, the entire experience must meet our culture standards. We encourage everyone to assist in creating a welcoming and safe environment. Please report<\/a> any concerns, harassing behavior, or suspicious or disruptive activity. Microsoft reserves the right to ask attendees to leave at any time at its sole discretion.\r\n
<\/div>\r\n
[msr-button text=\"Report a concern\" url=\"https:\/\/app.convercent.com\/en-us\/Anonymous\/IssueIntake\/LandingPage\/65d3b907-0933-e611-8105-000d3ab03673\" new-window=\"true\" ]<\/div>"},{"id":1,"name":"Agenda","content":"

Tuesday, April 13<\/h2>\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n
Time (EST)<\/strong><\/td>\r\nSession<\/strong><\/td>\r\n<\/td>\r\n<\/tr>\r\n
12:30 PM-12:35 PM<\/td>\r\nWelcome<\/strong>\r\nJessica Mastronardi, Senior Manager Community Engagement & Partnerships, Microsoft Research Outreach\r\nSusan Dumais, Technical Fellow and Lab Director, Microsoft Research Montreal, New England, and NYC<\/td>\r\n<\/td>\r\n<\/tr>\r\n
12:35 PM-1:05 PM<\/td>\r\nInteractive Representation Learning<\/strong>\r\nDevon Hjelm, Principal Researcher, Microsoft Research Montreal<\/td>\r\n<\/td>\r\n<\/tr>\r\n
1:05 PM-1:35 PM<\/td>\r\nChatbots: There and Back Again<\/strong>\r\nMahmoud Adada, Principal Engineer Manager, Turing Montreal<\/td>\r\n<\/td>\r\n<\/tr>\r\n
1:35 PM-2:05 PM<\/td>\r\nMachines with Common Sense? Modelling Plausibility in Text<\/strong>\r\nJackie Cheung, Assistant Professor, McGill University, Mila and Consulting Researcher, Microsoft Research Montreal<\/td>\r\n<\/td>\r\n<\/tr>\r\n
2:05 PM-2:15 PM<\/td>\r\n10 Minute Break<\/td>\r\n<\/td>\r\n<\/tr>\r\n
2:15 PM-3:25 PM<\/td>\r\nBreakout Session A: Generalization Failures in Deep Nets<\/strong>\r\nAlessandro Sordoni, Principal Researcher, Microsoft Research\r\nSara Hooker, Google Brain\r\nGauthier Gidel, Universit\u00e9 de Montr\u00e9al and Mila\r\nRemi Tachet des Combes, Microsoft Research\r\nAaron Courville, Universit\u00e9 de Montr\u00e9al and Mila<\/td>\r\nBreakout Session B: FATE concerns in NLP<\/strong>\r\nAlexandra Olteanu, Principal Researcher, Microsoft Research\r\nSu Lin Blodgett, Post Doc Researcher, Microsoft Research\r\nHal Daume III, Microsoft Research\r\nSolon Barocas, Microsoft Research\r\nJackie Cheung, Mcgill University and Mila\r\nAmandalynne Paullada, University of Washington<\/td>\r\n<\/tr>\r\n
3:25 PM-3:55 PM<\/td>\r\nDeep Learning of Abstract Mechanisms, Abstract Entities and Abstract Actions<\/strong>\r\nProfessor Yoshua Bengio, Founder and Scientific Director of Mila<\/td>\r\n<\/td>\r\n<\/tr>\r\n
3:55 PM-4:00 PM<\/td>\r\nClosing Discussion and Thank You Remarks<\/strong>\r\nProfessor Yoshua Bengio, Founder and Scientific Director of Mila\r\nPeter Lee, CVP, Research and Incubations, Microsoft Research<\/td>\r\n<\/td>\r\n<\/tr>\r\n<\/tbody>\r\n<\/table>\r\n
<\/div>\r\n

Wednesday, April 14<\/h2>\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n
Time (EST)<\/strong><\/td>\r\nSession<\/strong><\/td>\r\n<\/td>\r\n<\/tr>\r\n
12:30 PM-12:35 PM<\/td>\r\nWelcome<\/strong>\r\nJessica Mastronardi, Senior Manager Community Engagement & Partnerships, Microsoft Research Outreach\r\nSusan Dumais, Technical Fellow and Lab Director, Microsoft Research Montreal, New England, and NYC<\/td>\r\n<\/td>\r\n<\/tr>\r\n
12:35 PM-1:05 PM<\/td>\r\nTowards Hyperparameter-Free Optimization for Modern Machine Learning<\/strong>\r\nSimon Lacoste-Julien, Associate Professor, Universit\u00e9 de Montr\u00e9al, Mila<\/td>\r\n<\/td>\r\n<\/tr>\r\n
1:05 PM-1:35 PM<\/td>\r\nSuccessor Feature Sets for Planning and Imitation<\/strong>\r\nGeoff Gordon, Partner Researcher, Microsoft Research Montreal<\/td>\r\n<\/td>\r\n<\/tr>\r\n
1:35 PM-2:05 PM<\/td>\r\nDialogue as Collaborative Construction of Dataflow<\/strong>\r\nJason Eisner, Partner Principal Researcher, Microsoft Semantic Machines<\/td>\r\n<\/td>\r\n<\/tr>\r\n
2:05 PM-2:15 PM<\/td>\r\n10 Minute Break<\/td>\r\n<\/td>\r\n<\/tr>\r\n
2:15 PM-3:25 PM<\/td>\r\nBreakout Session A: Memory in RL<\/strong>\r\nIda Momennejad, Senior Researcher, Microsoft Research\r\nBlake Richards, Assistant Professor, McGill University, Mila\r\nGuillaume Lajoie, Assistant Professor, Universit\u00e9 de Montr\u00e9al, Mila<\/td>\r\nBreakout Session B: Observe, Remember, and Assist<\/strong>\r\nAndrew McNamara, Director of Engineering, Turing\r\nGeoff Gordon, Microsoft Research\r\nGerald Haslhofer, Microsoft (MMX)\r\nKaheer Suleman, Microsoft Research\r\nTim Franklin, Microsoft (MMX)\r\nTianyu Li, McGill University<\/td>\r\n<\/tr>\r\n
3:25 PM-3:55 PM<\/td>\r\nTowards Broad and Robust AI<\/strong>\r\nIrina Rish, Associate Professor, University of Montreal, Mila<\/td>\r\n<\/td>\r\n<\/tr>\r\n
3:55 PM-4:00 PM<\/td>\r\nClosing Discussion and Thank You Remarks<\/strong>\r\nProfessor Yoshua Bengio, Founder and Scientific Director of Mila\r\nPeter Lee, CVP, Research and Incubations, Microsoft Research<\/td>\r\n<\/td>\r\n<\/tr>\r\n<\/tbody>\r\n<\/table>"},{"id":2,"name":"Breakout Sessions","content":"

Tuesday, April 13 | 2:15 PM - 3:25 PM EST<\/h2>\r\nOne the day of the workshop, select one of the following breakout sessions to attend.\r\n

Breakout Session A: Generalization Failures in Deep Nets<\/strong><\/h3>\r\nFacilitator:\r\nAlessandro Sordoni, Microsoft Research Montreal\r\nRemi Tachet des Combes, Microsoft Research Montreal\r\n\r\nPanelists:\r\nSara Hooker, Google Brain\r\nGauthier Gidel, Universit\u00e9 de Montr\u00e9al and Mila\r\nAaron Courville, Universit\u00e9 de Montr\u00e9al and Mila\r\n\r\nAbstract:\r\nAlthough large (overparametrized) deep neural networks do wonders, level-5 autonomous driving appears still out of reach. In this short and speculative session, we won't brainstorm about autonomous driving (yet), but instead will question how deep nets behave in the long tail, on adversarial examples, with spurious correlations and more generally out-of-distribution. We will hear about some generalization failures of deep nets as well as recent approaches that have been explored to fix them. This breakout is intended to be for both Neural Nets lovers and skeptics\r\n\r\nThis session will start with a panel discussion, followed by a live Q&A with participants.\r\n

Breakout Session B: FATE Concerns in NLP<\/strong><\/h3>\r\nFacilitators:\r\nAlexandra Olteanu, Microsoft Research Montreal\r\nSu Lin Blodgett, Microsoft Research Montreal\r\n\r\nPanelists:\r\nAmandalynne Paullada, University of Washington\r\nHal Daume III, Microsoft Research New York City\r\nJackie Cheung, McGill University and Mila\r\nSolon Barocas, Microsoft Research New York City\r\n\r\nAbstract:\r\nIn this breakout session we will discuss existing concerns related to the fairness, accountability, transparency and ethics (FATE) of language technologies (and the data they ingest or generate), as these technologies might surface, replicate, exacerbate or even cause a range of computational harms (from surfacing offensive speech or reinforcing stereotypes, to even more subtle issues like nudging users towards undesirable patterns of behavior or triggering memories of traumatic events). How can we reliably measure fairness related harms? Whose data is included when training the model? How do we better foresee potential harms? How do we engage stakeholders? And many other critical questions.\r\n\r\nThis session will start with a panel discussion, followed by a live Q&A with participants.\r\n

Wednesday, April 14 | 2:15 PM - 3:25 PM EST<\/h2>\r\nOne the day of the workshop, select one of the following breakout sessions to attend.\r\n

Breakout Session A: Memory in RL<\/strong><\/h3>\r\nFacilitators:\r\nBlake Richards, McGill University and Mila\r\nGuillaume LaJoie, Universit\u00e9 de Montr\u00e9al and Mila\r\nIda Momennejad, Microsoft Research NYC\r\n\r\nAbstract:\r\nThe early development of reinforcement learning (RL) was directly inspired by ideas from neuroscience and psychology, but over time, RL produced theories and techniques that bore less and less resemblance to mechanisms in the brain. However, there has been growing interest in the RL community to re-examine the potential for insights about memory in the brain to inform RL. In particular, the brain's declarative memory systems (episodic and semantic) may provide guidance as to how to build RL systems that can avoid catastrophic forgetting, engage in long-distance credit assignment, and generalize to novel situations. In this workshop we will discuss the work to-date on memory systems for RL, some new insights into memory from neuroscience, and potential future directions in this area.\r\n\r\nThis session will start with a topic overview, followed by a live discussion and Q&A with participants.\r\n

Breakout Session B: Observe, Remember, and Assist<\/strong><\/h3>\r\nFacilitator:\r\nAndrew McNamara, Turing Montreal\r\n\r\nSpeakers:\r\nTalk 1: Observe Remember Automation: AI-Assisted Productivity\r\nTim Franklin, Microsoft\r\n\r\nTalk 2: Automating Style Transfer in Office Documents\r\nTianyu Li, McGill University\r\n\r\nPanelists:\r\nGeoff Gordon, Microsoft Research\r\nGerald Haslhofer, Microsoft\r\nKaheer Suleman, Microsoft Research\r\n\r\nAbstract:\r\nObserve, Remember, and Automate is a formulation of remote process automation whereby we seek to empower users to achieve more. Our goal is to automate as much as possible for users to limit their time on mundane and repetitive tasks. We want to help users optimize their time by making recommendations and predictions based on real time context and actions taking into account everything we've observed and remembered from past interactions. This effort requires the combination of various research areas including language modeling, representation learning, computer vision, reinforcement learning, graph representations, and much more.\r\n\r\nThis session will start with a topic overview, followed by a live discussion and Q&A with participants."}],"msr_startdate":"","msr_enddate":"","msr_event_time":"","msr_location":"Virtual (Microsoft Teams)","msr_event_link":"","msr_event_recording_link":"","msr_startdate_formatted":"","msr_register_text":"Register now","msr_cta_link":"","msr_cta_text":"","msr_cta_bi_name":"","featured_image_thumbnail":"\"MSR","event_excerpt":"You\u2019re invited to join us for the 2021 MSR & Mila Research Workshop held virtually on Microsoft Teams. The workshop will feature presentations and forward-thinking breakout sessions by Mila professors, students and researchers at Microsoft. The workshop is an opportunity for Mila professors, students and MSR researchers to share and present what they\u2019re currently working on with the goal of facilitating and building on existing research collaborations between Mila & MSR. Organizing Committee Brittney Muller,…","msr_research_lab":[437514],"related-researchers":[],"msr_impact_theme":[],"related-academic-programs":[],"related-groups":[],"related-projects":[],"related-opportunities":[],"related-publications":[],"related-videos":[741427,741445],"related-posts":[],"_links":{"self":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-event\/712540","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-event"}],"about":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/types\/msr-event"}],"version-history":[{"count":2,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-event\/712540\/revisions"}],"predecessor-version":[{"id":1146885,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-event\/712540\/revisions\/1146885"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/media\/718087"}],"wp:attachment":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/media?parent=712540"}],"wp:term":[{"taxonomy":"msr-research-area","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/research-area?post=712540"},{"taxonomy":"msr-region","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-region?post=712540"},{"taxonomy":"msr-event-type","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-event-type?post=712540"},{"taxonomy":"msr-video-type","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-video-type?post=712540"},{"taxonomy":"msr-locale","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-locale?post=712540"},{"taxonomy":"msr-program-audience","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-program-audience?post=712540"},{"taxonomy":"msr-post-option","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-post-option?post=712540"},{"taxonomy":"msr-impact-theme","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-impact-theme?post=712540"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}