Uni[MASK]: Unified Inference in Sequential Decision Problems
Micah Carroll / UC Berkeley
Feb 28, 2023
Randomly masking and predicting word tokens has been a successful approach in pre-training language models for a variety of downstream tasks. In this work, we observe that the same idea also...
Quality Diversity Scenario Generation for Robust Autonomy
Matthew Fontaine / University of Southern California
Sep 15, 2022
As artificial agents and robots become more advanced through modern machine learning and complex symbolic methods, testing these systems becomes increasingly difficult. Trained agents must be...
Novel Opportunities in Open-Endedness
Franziska Meier / Meta AI
Aug 24, 2022
As interest in the field of open-endedness expands, ideas like continual discovery and increasingly complexity have gained significant attention. The aim of this talk is to bring attention to two...
Lifelong Learning for Robotics
Franziska Meier / Meta AI
March 24, 2022
While there has been major investment in developing robot learning algorithms, achieving true autonomy remains a wide-open research question. A key limitation of current learning approaches is the...
Understanding the World Through Action
Sergey Levine / UC Berkeley
January 6, 2022
The capabilities of modern machine learning systems are to a large extent determined by their ability to effectively utilize large and diverse datasets. However, such systems typically focus on...
Learning Structured Models of the World
Thomas Kipf / Google Brain
December 16, 2021
The world around us — and our understanding of it — is rich in compositional structure: from atoms and their interactions to objects and entities in our environments. How can we learn models of...
Open-Ended Learning Leads to Generally Capable Agents
Max Jaderberg / DeepMind
November 25, 2021
In this talk I will cover our recent publication "Open-Ended Learning Leads to Generally Capable Agents". In this work we turn our attention to how to create embodied agents in simulation that can...
Training Virtual Robots in Realistic Simulators
Erik Wijmans / Georgia Institute of Technology
September 2, 2021
Recently there has been a shift in computer vision research from static vision, e.g., object detection, to Embodied AI, e.g., robot navigation. In this talk, I will focus on the task of PointGoal...
Towards Human-Like And Collaborative AI in Video Games
Katja Hofmann / Microsoft Research
August 19, 2021
Developing agents capable of learning complex human-like behaviors is a key goal of artificial intelligence research. Progress towards this goal has exciting potential for applications in video...
Towards Knowledge Grounded Text Agents
Matthew Hausknecht / Microsoft Research
July 8, 2021
Learning agents in text-based games are faced with the challenge of understanding and generating language to accomplish various goals. Decomposing this primary challenge, I will highlight recent...
Using Video Games To Reverse Engineer Human Intelligence
Sam Gershman / Harvard University
June 21, 2021
Video games have become an attractive testbed for evaluating AI systems, by capturing some aspects of real-world complexity (rich visual stimuli and non-trivial decision policies) while...
Exploring Context for Better Generalization in Reinforcement Learning
Amy Zhang / UC Berkeley and Facebook AI Research
June 7, 2021
The benefit of multi-task learning over single-task learning relies on the ability to use relations across tasks to improve performance on any single task. While sharing representations is an...
Some Reasons To Do Embodied Machine Learning
Felix Hill / DeepMind
May 24, 2021
In this talk, I'll give four good reasons to do embodied machine learning. Learning in an agent that can perceive and interact with its environment is fundamentally different from other ML...
Mental Simulation, Imagination, and Model-Based Deep RL
Jessica Hamrick / DeepMind
May 10, 2021
Mental simulation—the capacity to imagine what will or what could be—is a salient feature of human cognition, playing a key role in a wide range of cognitive abilities. In artificial intelligence,...
Towards multi-agent emergent communication as a building block of human-centric AI
Angeliki Lazaridou / DeepMind
Apr 26, 2021
The ability to cooperate through language is a defining feature of humans. As the perceptual, motory and planning capabilities of deep artificial networks increase, researchers are studying whether they can also develop a shared language to interact. In this talk, I will highlight recent advances in this field but also common headaches (or perhaps limitations) with respect to experimental setup and evaluation of emergent communication. Towards making multi-agent communication a building block of human-centric AI, and by drawing from my own recent work, I will discuss approaches on making...
Finding Good Representation for Search and Exploration in RL
Yuandong Tian / Facebook AI Research
Apr 12, 2021
How to learn good latent representations is an important topic in the modern era of machine learning. For reinforcement learning, using a good representation makes the decision-making process much...
Evolutionary Algorithms and Game AI
Simon Lucas / Queen Mary University of London
Mar 29, 2021
Evolutionary algorithms are powerful black-box optimisers that find many applications in Game AI. They can be applied in real-time to provide robust policies across a range of games, or at design...
On Question Answering on Images and Databases
Dzmitry Bahdanau / McGill University
Mar 15, 2021
The AI's ability to answer questions that are grounded in context is interesting from both academic and practical perspectives. In this talk, I will present two projects that study question answering (QA) in visual and symbolic database context respectively. In the first project my collaborators and I show that a neuro-symbolic visual QA system can learn variable bindings from the top-down QA training signal only. The system successfully learns both conventional bindings (e.g. objects) and less trivial ones (e.g. groups). In our second project my colleagues at Element AI and I explore...
Learning to Cooperate, Communicate and Coordinate
Jakob Foerster / Facebook AI Research
Mar 08, 2021
In recent years we have seen rapid progress on a number of zero-sum benchmark problems in artificial intelligence, e.g. Go, Poker and Dota. In contrast to these competitive settings, success in...
Grounded Language Learning Without Grounded Supervision
Jacob Andreas / MIT
Mar 01, 2021
Central to tasks like instruction following and question answering is the ability to ground linguistic understanding in perception and action. Machine learning models for these tasks typically...
Increasing generality in reinforcement learning through procedural content generation (or have fun trying)
Julian Togelius / New York University
Jan 18, 2021
Julian Togelius will be giving a talk with the title: "Increasing generality in reinforcement learning through procedural content generation (or have fun trying)", exploring how to make...
Intuitive Reasoning as (Un)supervised Neural Generation
Yejin Choi / University of Washington
Jan 04, 2021
Neural language models, as they grow in scale, continue to surprise us with utterly nonsensical and counterintuitive errors despite their otherwise remarkable performances on leaderboards. In this...
Social Reinforcement Learning
Natasha Jaques / Google Brain
Dec 21, 2020
Social learning helps humans and animals rapidly adapt to new circumstances, and drives the emergence of complex learned behaviors. This talk focuses on Social Reinforcement Learning, developing...
Promise, Progress, and Challenges in Open-Ended Machine Learning
Joel Lehman / OpenAI
Nov 23, 2020
Researchers in open-ended machine learning are inspired by natural and human processes of innovation (like biological evolution or science itself), and aim to uncover the engineering principles underlying boundlessly creative search algorithms, i.e. algorithms capable of continual production of useful and interesting innovations, The speculative promise for machine learning from such open-ended search includes automated discovery of curricula for reinforcement learning, new neural architectures, and most ambitiously, AGI itself. While its most ambitious potential is far from being...