(Track1) Advances in Approximate Inference Q&A
Women in Machine Learning
-----
To help navigate the virtual WiML Workshop @ NeurIPS 2020, we have compiled a list of important links and guidelines below:
1. All activities, such as the talks, mentorship roundtables, sponsor expo, and social events, will take place in the Gather.Town WiML world ([ protected link dropped ] Please fill out the participation form and accept the WiML Code of Conduct to access the Gather.Town: [ protected link dropped ] *Please note that the poster session will take place on Monday, December 7th at 8:30-10:30p UTC / 3:30-5:30p EST / 12:30-2:30p PST.*
3. Asking questions in live Q&A: To ask the speaker a question, you may: (a) join Zoom by clicking the link at the top of this page; or (b) type your question directly into the Rocket.Chat at the top of this page. The moderator will monitor the chat to convey selected questions to the speaker.
4. Please refer to the WiML workshop program book for details on all the program activities: https://bit.ly/3mOgpXE
5. The Gather.Town guidelines contains information about how to make full use of your experience at Gather.Town as well as troubleshooting common technical issues: https://bit.ly/2Igp45Z
6. Finally, we’d love to hear your thoughts on the virtual workshop experience. Please fill out the exit survey at the end of the workshop and enter for a chance to win a $$25 gift card: bit.ly/3m3v0ht
(Track2) Explaining Machine Learning Predictions: State-of-the-art, Challenges, and Opportunities Q&A
The Real AI Revolution
The two long-held aspirations to understand the mechanisms of human intelligence, and to recreate such intelligence in machines, have inspired many of us to build our careers in the field of machine learning. However, while the creation of technologies supporting general intelligence would be truly revolutionary, such an achievement still seems to lie well into the future. Meanwhile, another profound revolution, also built on machine learning, is already unfolding and is set to transform almost every aspect of our lives. In this talk I will highlight the nature of this revolution and why the coming decade will be a hugely exciting, and critically important, time to engage deeply in machine learning for those who want to have a truly transformational impact in the real world.
Orals & Spotlights Track 21: Optimization
Orals & Spotlights Track 19: Probabilistic/Causality
Orals & Spotlights Track 20: Social/Adversarial Learning
Orals & Spotlights Track 17: Kernel Methods/Optimization
Orals & Spotlights Track 16: Continual/Meta/Misc Learning
Orals & Spotlights Track 18: Deep Learning
Orals & Spotlights Track 15: COVID/Applications/Composition
In Memory of Olivier Chapelle
Earlier this year, Olivier Chapelle passed away aged 42, after an unexpected severe illness that he was diagnosed with only a year before. He is sorely missed by his family and his friends, many of which had known him not only as a kind and modest human being, but also as one of the very best machine learning researchers of his generation. We are holding a memorial session in Olivier’s honour, with contributions from some of these friends, including Olivier Bousquet, Andre Elisseeff, Minmin Chen, Dilan Görür, Isabelle Guyon, Bernhard Schölkopf, Keerthi Selvaraj, Vladimir Vapnik, Kilian Weinberger, and Jaston Weston.
COVID-19 Symposium Day 2
The COVID-19 global pandemic has disrupted nearly all aspects of modern life. This year NeurIPS will host a symposium on COVID-19 to frame challenges and opportunities for the machine learning community and to foster a frank discussion on the role of machine learning. A central focus of this symposium will be clearly outlining key areas where machine learning is and is not likely to make a substantive impact. The one-day event will feature talks from leading epidemiologists, biotech leaders, policy makers, and global health experts. Attendees of this symposium will gain a deeper understanding of the current state of the COVID-19 pandemic, challenges and limitations for current machine learning capabilities, how machine learning is accelerating COVID-19 vaccine development, and possible ways machine learning may aid in the present and future pandemics.
In ML we trust: the Trustworthy ML Social
We propose a 2-hour social with the following activities:
The first hour will consist of 20-30 min participant-driven breakout sessions centered around different aspects of trustworthy ML (e.g. fairness, interpretability, robustness, etc.), with the goal of curating 1-2 questions on each topic. This will be followed by a 40-min panel discussion where panelists address the participant-curated questions.
The second hour will consist of trivia games with both fun and technical questions. We will design some games with a trustworthy ML spin, e.g. trustworthy scrabble, roulette, charades, etc., to promote educational interaction in a casual environment.
If there is time left we will ask participants to return to the breakout rooms for socializing.
- The first hour will consist of 20-30 min participant-driven breakout sessions centered around different aspects of trustworthy ML (e.g. fairness, interpretability, robustness, etc.), with the goal of curating 1-2 questions on each topic. This will be followed by a 40-min panel discussion where panelists address the participant-curated questions.
- The second hour will consist of trivia games with both fun and technical questions. We will design some games with a trustworthy ML spin, e.g. trustworthy scrabble, roulette, charades, etc., to promote educational interaction in a casual environment.
- If there is time left we will ask participants to return to the breakout rooms for socializing.
Computational Biology Speed Networking
Join us for computational biology speed networking! We will run two one-hour sessions. During each hour, participants will be randomly places in breakout rooms of 3 - 4 people for 6-minute mini-meetings. We will provide suggested questions and discussion topics, but participants are also free to choose their own.
(Track2) Practical Uncertainty Estimation and Out-of-Distribution Robustness in Deep Learning Q&A
Deep learning models are bad at signalling failure: They tend to make predictions with high confidence, and this is problematic in real-world applications such as healthcare, self-driving cars, and natural language systems, where there are considerable safety implications, or where there are discrepancies between the training data and data that the model makes predictions on. There is a pressing need both for understanding when models should not make predictions and improving model robustness to natural changes in the data.
This tutorial will give an overview of the landscape of uncertainty and robustness in deep learning. Namely, we examine calibration and out-of-distribution generalization as key tasks. Then we will go into a deep dive into promising avenues. This includes methods which average over multiple neural network predictions such as Bayesian neural nets, ensembles, and Gaussian processes; methods on the frontier of scale in terms of their overall parameter or prediction-time efficiency; and methods which encourage key inductive biases such as data augmentation. We ground these ideas in both empirical understanding and theory, and we provide practical recommendations with baselines and tips & tricks. Finally, we highlight open challenges in the field.
Friendly "Super Smash Brothers Melee" Tournament
An online tournament of "Super Smash Brothers Melee", welcoming players of all skill levels.
Thanks to recent community effort, online play is virtually lag-free: https://twitter.com/Fizzi36/status/1275096470765490176
We will likely hold the tournament in gather.town for proximity chat.
We may casually encourage "academic speed-dating" throughout. For example, we might suggest that players discuss research with each other after their match is done, but before their next match begins.
If the players decide to share their screens, spectators will be able to watch the game or join in discussion.
As the tournament progresses, people will be free to play against others in a non-competitive environment.
Catalyzing the Collaborative Development of AI
Informal and interactive session on how we can work together to make the development of AI more collaborative within the research community.
We will present Stateoftheart AI, a new, free, and open-data platform that aims to facilitate this collaboration by mapping and visualizing several aspects of AI; as well as creating a repository of models and datasets aiming for compatibility and standardization.
Open to all. For the benefit of society.
Race After Technology
Let’s come together for Un-Bookclub Race After Technology social at NeurIPS. We’ve been learning a lot from the discussions in the book clubs we’ve been running out of the book ‘Race After Technology: Abolitionist Tools for the New Jim Code’ by Dr. Ruha Benjamin.
We got our first copy of the book as a gift at the amazing Black in in AI dinner at NeurIPS 2019 where Dr. Benjamin spoke. We’d love to give you the gift of connection, conversation, and reflection Dr. Benjamin gave us.
Prework: please read the book or watch ICLR keynote: 2020 Vision: Reimagining the Default Settings of Technology & Society, Prof. Ruha Benjamin / Princeton: https://iclr.cc/virtual2020/speaker3.html
AI for Geospatial Social
Geospatial data is vital to understanding our world and humanity’s place in it. The latest Artificial Intelligence techniques form a crucial component in turning geospatial data into information that can be used to combat climate change, respond to natural disasters, protect endangered species, grow better crops, protect our oceans, and revolutionize transport. The AI for Geospatial social serves as a way for attendees from different fields that may not interact during the workshops or main sessions to get to know one another and collaborate in an informal setting. Themed breakout rooms and icebreakers (trivia sessions with prizes) will facilitate an engaging experience and provide an opportunity to learn something both about geospatial analytics and some of its lesser known pioneers.
(Track1) Abstraction & Reasoning in AI systems: Modern Perspectives Q&A
A Future of Work for the Invisible Workers in A.I.
The A.I. industry has created new jobs that have been essential to the real-world deployment of intelligent systems. These new jobs typically focus on labeling data for machine learning models or having workers complete tasks that A.I. alone cannot do. Human labor with A.I. has powered a futuristic reality where self-driving cars and voice assistants are now commonplace. However, the workers powering our A.I. industry are often invisible to consumers. Together, this has facilitated a reality where these invisible workers are often paid below minimum wage and have limited career growth opportunities. In this talk, I will present how we can design a future of work for empowering the invisible workers behind our A.I. I propose a framework that transforms invisible A.I. labor into opportunities for skill growth, hourly wage increase, and facilitates transitioning to new creative jobs that are unlikely to be automated in the future. Taking inspiration from social theories on solidarity and collective action, my framework introduces two new techniques for creating career ladders within invisible A.I. labor: a) Solidarity Blockers, computational methods that use solidarity to collectively organize workers to help each other to build new skills while completing invisible labor; and b) Entrepreneur Blocks, computational techniques that, inspired from collective action theory, guide invisible workers to create new creative solutions and startups in their communities. I will present case-studies showcasing how this framework can drive positive social change for the invisible workers in our A.I. industry. I will also connect how governments and civic organizations in Latin America and U.S. rural states can use the proposed framework to provide new and fair job opportunities. In contrast to prior research that focused primarily on improving A.I., this talk will empower you to create a future that has solidarity with the invisible workers in our A.I. industry.
