Socials

Ehi Nosakhare, Tatjana Chavdarova

WiML is excited to announce a Virtual Social @ ICLR 2021 involving a Virtual Panel and a Socializing Session. During the Virtual Panel, we hope to discuss how we can start and navigate careers through the COVID-19 pandemic. The panel features ML researchers at various career stages who will talk about their experience networking, job hunting, collaborating and/or starting a new position in a primarily online environment. After the panel, we will adjourn to the icebreaker.video platform for socializing.

Our panelists are:

  • Candace Ross (PhD student in Computer Science, MIT)

  • Christina Papadimitriou (Machine Learning Engineer, JPMorgan Chase)

  • Claire Vernade (Research Scientist, DeepMind, London)

  • Po-Ling Loh (Lecturer in the Department of Pure Mathematics and Mathematical Statistics, University of Cambridge)

  • Sinead Williamson (Assistant Professor of Statistics, University of Texas at Austin)

The panel will be moderated by Ehi Nosakhare (Senior Data and Applied Science Manager, Microsoft).

Everyone registered for ICLR is encouraged to attend! We expect all attendees to adhere to the WiML Code of Conduct: https://wimlworkshop.org/conduct/.

Jacob Beck, David Abel

The event will be a panel and discussion centered around AGI and philosophy, divided into two days.

On May 3rd, we will discuss more tractable machine learning questions such as “How do we work toward AGI” and “How do current RL frameworks most drastically differ from the biologic and evolutionary framework that gave rise to human intelligence?” On May 6th, we will discuss philosophical questions such as “Is a simulated brain conscious” and “Should we work toward AGI”. Each day will consist of discussion between the invited panelists, followed by questions and “breakout rooms” for more casual discussion.

Panelists on May 3rd will include: Michael Littman, Melanie Mitchell, Jane Wang, Matt Botvinick

Panelists on May 6th will include Nick Bostrom, Christopher Hill

Ana Lucic, Rosie Campbell

Although considerable progress has been made on utilizing AI for healthcare, in both personal and public health, there still exists a gap between research and practice when operationalizing it. What are the challenges in operationalizing AI in healthcare applications such as clinical diagnosis, drug development, patient triaging, individual treatment recommendations, epidemic forecasting, and outbreak detection? How do these challenges differ across various regions in the world? In this social, we want to exchange ideas for how to overcome these challenges by connecting researchers and practitioners from within the ICLR community to share their experiences with (attempting to) solve problems in healthcare with AI. We aim to foster a community of stakeholders involved in this space, including researchers, practitioners, and affected groups, in order to understand the needs and values of all stakeholders involved in the process.

Jung-Woo Ha, Gunhee Kim

We invite everyone who is part of and/or interested in the ML research scene in Korea. Participants can introduce their own ML research, especially if it's part of ICLR 2021. They can also introduce ICLR 2021 papers that they find interesting and discuss them with other participants. Other possible discussion topics include (but are not limited to): Korean NLP, computer vision and datasets, ML research for COVID-19, ML for post COVID-19 era, and career options in academia/industry in Korea. We welcome everyone from anywhere in the world, as long as you can keep awake if our event falls in the middle of the night for you.

Tahiya Chowdhury, Shiran Dudy

This social will focus on a discussion on finding a balance between the good and the ugly side of AI research. On one side, AI promises automation and availability that can relieve humans from mundane tasks. On the other, the power of AI in society comes with the price of systematic bias, misinformation, and disruption in the labor market. In this fast approaching field of research, how to find meaning in our research that can serve as a moral compass so that we don’t lose focus on problems that can make a true positive impact on society? The social will be interactive where participants will share their thoughts and experiences on societal impact, public opinion of AI, and AI ethics in connection to finding purpose in research and works on AI.

Vaidheeswaran Archana, Soham Chatterjee Gopal

The covid vaccine administration is one of the biggest globalisation efforts of this century. With less than 2% of the entire world population vaccinated, it is important now more than ever to understand the demographic makeup in covid vaccine administration. For instance, privatisation and a lack of standardisation in the manufacturing and distribution of covid vaccines has caused a pattern where richer countries have been able to put more resources into the procurement of vaccine doses for their citizens (https://globalhealth.duke.edu/news/if-rich-countries-dont-share-their-vaccines-pandemic-could-stretch-years). Recent data released by the CDC has also shown that vaccines are predominantly being administered in the white, non-hispanic population (https://covid.cdc.gov/covid-data-tracker/#vaccination-demographic).

In this virtual hall, we would like to invite participants to put forth their findings as well as discuss how to gather and analyse data to understand the role that demographics is playing in this global vaccination drive. Finally, by analysing this data, we can predict how vulnerable communities are affected during pandemics. This can serve as important findings for public health organisations to deal with future pandemics.

Jennifer Hobbs, Sujoy Ganguly

“Lapsed” (aka. Former) Physicists are plentiful in the machine learning community. Inspired by Wine and Cheese seminars at many institutions, this BYOB event is an informal opportunity to connect with members of the community. Hear how others made the transition between fields, discuss how your physics training prepared you to switch fields, what synergies between physics and machine learning excite you the most, share your favorite physics jokes your computer science colleagues don’t get, and just meet other cool people. Open to everyone, not just physicists, you just have to tolerate our humor. Wine and Cheese encouraged, but not required.

Anil Goyal, Sudarsun Santhiappan, Jeshuren Chelladurai Ram Swaminathan

With the explosion of electronic data in the healthcare space, the industry needs to find the best ways to extract relevant information from the data to help various stakeholders, such as doctors, patients, hospitals, and insurance companies. In this social event, we would be discussing various challenges while dealing with real-world unstructured healthcare data and how BUDDI.AI uses ML, NLP tech to solve these challenges. Further, we would also discuss about how the extracted information is used in various use-cases like Medical Autocoding, Pharma covigilance, Clinical Decision Systems, Claim Denial Prediction, Clinical Trials, Clinical Documentation Improvement and Revenue Cycle Management.

Tijmen Blankevoort

Our presentations are most likely the highest impact activities we have as researchers. They are oftentimes quite dense. In those 10 minutes in your conference oral, you have the chance to show your work to a large audience for world-wide recognition. This is both incredibly stressful and difficult to do. The months of research that you've done, with all the ideas and all the results, have to be jam packed in a short time interval, and your audience is tired of the long conference and the information hoses they are drinking from.

How do you make the most out of your presentation? How do you make sure that people understand your work, get excited by it, and remember you in the future? In this session we will cover: - How to structure your presentation and storytelling - Captivating your audience and making them remember you - Guiding your audience through tough and difficult to parse material - Dynamic and easy to follow slide creation - Preparation for the big moment - Frequent mistakes and the psychology of insecurity.

Your teacher for the session will be Tijmen Blankevoort. A public speaker with over 7 years of experience, recurring radio and podcast …

Braham Snyder, Alex Lewandowski

Friendly matches of "Super Smash Brothers Melee", welcoming players of all skill levels.

For setup see https://slippi.gg/netplay — online play is virtually lag-free.

We casually encourage everyone to discuss research with each other between matches.

Stephanie Fong, Nazneen Rajani

Taking the virtual place of the "water coolers" throughout the in-person Conference floor typically - this social will be a casual open forum to discuss topics pertaining to being a WiAI for both womxn and allies, and include topics around experience, challenges, collaboration, and advice for others. Join us to share, network or just listen in!

Madhulika Srikumar, Rosie Campbell

There is an emerging consensus that AI researchers and scientists must consider the societal impact of their research. This idea is already taking hold with the introduction of Broader Impact Statements at NeurIPS last year and Paper Checklist Guidelines for authors at this year’s edition. This social will explore the practical challenges that arise for researchers with the introduction of these new requirements. The social will provide the ICLR community with the opportunity to engage with researchers who are thinking about these questions in an informal setting. Come armed with questions, find peers that are facing similar challenges in their research and learn more about resources for anticipating downstream consequences!

There will be two sessions to accommodate researchers from different time zones. Please go to gather.town.

Taylor, Ian

Not everyone makes their mark in data science through the path of traditional education. Join us to share your unique background story into data science; and hear from founder and CEO of Grid.ai, William Falcon, PyTorch Lightning's Head of Developer Advocacy, Ari Bornstein, and AI Influencer Aishwarya Srinivasan on their untraditional paths. We hope to see you there!

Tijmen Blankevoort, Marian Verhelst, Daniel Soudry, Max Welling

Panelists:

Max Welling (University of Amsterdam, Qualcomm), Marian Verhelst (KU Leuven), Daniel Soudry (Technion University)

Session Summary:

According to recent research by University of Massachusetts at Amherst, the amount of power required for training a popular transformer neural network with 213 million parameters is equivalent to the emissions of roughly 626,000 pounds of carbon dioxide. That’s nearly five times the lifetime emissions of the average U.S. car, including its manufacturing. For that purpose, we turn our attention to research that successfully manages to quantize neural networks. We sit down virtually to discuss how we can make AI models smaller and more power- and energy-efficient. Our guests will discuss trends and opportunities in this field and then we’ll open the floor for questions and conversation with the audience.

William Agnew
Luisa Zintgraf, Jakob Foerster, Maximilian Igl, Christian Schroeder de Witt

Global challenges like climate change, human rights, and the ongoing pandemic urgently require international cooperation and coordination. Instead, the world is currently being destabilized by a new arms race fuelled by automation and remote controlled weapons like drones and robots. Recent advances in AI research have drastically reduced the technological barrier for the full automation of these weapons, blurring the line between armed drones and lethal autonomous weapon systems (LAWS). LAWS have been recognised as a substantial danger to humanity by many institutions and individuals in the AI community (see this open letter).

Despite considerable campaign efforts, the automation of warfare is progressing quickly, with disastrous consequences. We as AI researchers have the possibility to make a difference, as the military sector depends on civil science and we can make our voice heard in public. But we need to do more, and do it fast.

The aim of this Social is to provide information and discuss how we, as a community, can come together to prevent the further development and proliferation of LAWS. Therefore we also want to discuss the necessity of internationally controlling, disarming and banning remote controlled weapons, such as drones. The Social will consist of talks …

Daneil Chambers, Fatoumata Fall

We will explore and experience the facts and realities of the gender pay gap through Kahoot, an experiential activity, and breakout rooms of attendees networking and sharing stories. We will also hear from industry veteran Dr. Fanny Nina-Paravecino about the importance of negotiating for successful career advancement. Fanny will share stories as an experienced Research Scientist who is now a Tech manager.

Mah Parsa, Sepid Parsa

WinAIML aims to bring women that are passionate about Creating New AI, Enhancing Current AI, Applying Traditional AI together.

WinAIML is a group to allow women: 1. to learn from other women in the field of AI 2. to share their pieces of knowledge in the field of AI 3. to understand what is going on in the future of AI.

WinAIML aims to empower women to have an active role in the field of AI.

Laura Montoya, Maria Luisa Santiago, Dennis Núñez Fernández, Javier Turek, Sara Iris Garcia

This LatinX in AI (LXAI) social is aimed at LatinX individuals working on or interested in AI with a goal to increase the visibility of researchers of LatinX origin. Those already working in AI will have the opportunity to connect with fellow LatinX and make their own work known, while those new to the field will benefit from the scientific exchange, guidance, and advice of researchers with their same background. Participants will be able to engage in discussions about AI (formal and informal) and to share their thoughts on how to increase the presence of LatinX in AI.

During this social, we’ll feature invited talks from prominent LatinX in AI community members and host roundtable discussions facilitating conversations among attendees.

Yiyuan Li, Rishabh Joshi

Language research is tightly connected with and widely benefits from the advance of learning approaches, namely language modeling, few-shot learning, meta-learning approaches and representation learning. ‘ML and Language’ is the place to exchange ideas and opinions for the frontier and future of those directions, and raise attention to the challenges. We want to bring researchers together to share ideas, foster collaborations and discuss their problems specifically in ML applied to language. Come and also join us in the session and https://bit.ly/3esHwVk.

Laura Montoya, Maria Luisa Santiago, Dennis Núñez Fernández, Javier Turek, Sara Iris Garcia

This LatinX in AI (LXAI) social is aimed at LatinX individuals working on or interested in AI with a goal to increase the visibility of researchers of LatinX origin. Those already working in AI will have the opportunity to connect with fellow LatinX and make their own work known, while those new to the field will benefit from the scientific exchange, guidance, and advice of researchers with their same background. Participants will be able to engage in discussions about AI (formal and informal) and to share their thoughts on how to increase the presence of LatinX in AI.

During this social, we’ll feature invited talks from prominent LatinX in AI community members and host roundtable discussions facilitating conversations among attendees.

Alex Bezzubov, Timofey Bryksin

ML for Software Engineering Social brings together researchers working on applying ML to source code modeling and other software engineering (SE) tasks for building better developer tools.

We allocate some time at the beginning for short demos/presentations of ongoing work by different groups and spend the rest of the time on informal discussion on

  • What are the challenges of navigating research at the intersection of environments (academic and industrial research labs) and fields (ML, SE, PL)?

  • What are the frontiers of ML4SE research? What are the important unsolved problems in the area?

  • What are the biggest hurdles in creating a new ML for SE? (e.g. data, compute, etc)

  • Computer vision had ImageNet. NLP has [Super]GLUE. If it were up to you, what benchmark would you have the community work on?

  • What do the recent advances in NLP, such as Transformers, mean for the ML for SE area?

  • Common practical challenges of research results adoption in the industry.

Elise van der Pol, Karen Ullrich, Yuge Shi

Tired of rigorous scientific processes? We offer the fastest publication cycle ever with no review (!!) at the Bad Hypothesis Contest. This ICLR 2021 social is organised in the format of a contest. Each participant will prepare/improvise a 5 minutes stand-up, where they present a bad hypothesis (that is clearly ridiculous and untrue) and provide “evidence” (spurious correlations) that supports this hypothesis. We provide a list of hypotheses to choose from on the website (https://www.elisevanderpol.nl/badhypothesiscontest/), however everyone is welcomed to propose their own bad hypothesis. Our format is inspired by BAHfest.

Jacob Beck, David Abel, Clare Lyle

The event will be a panel and discussion centered around AGI and philosophy, divided into two days.

On May 3rd, we will discuss more tractable machine learning questions such as “How do we work toward AGI” and “How do current RL frameworks most drastically differ from the biologic and evolutionary framework that gave rise to human intelligence?” On May 6th, we will discuss philosophical questions such as “Is a simulated brain conscious” and “Should we work toward AGI”. Each day will consist of discussion between the invited panelists, followed by questions and “breakout rooms” for more casual discussion.

Panelists on May 3rd will include: Michael Littman, Melanie Mitchell, Jane Wang, Matt Botvinick

Panelists on May 6th will include Nick Bostrom, Christopher Hill

Rosanne Liu, Jason Yosinski, Brian Cheung, Janice Lan, Suzana Ilić, Ryan Teehan, Natalie Summers, Connor Leahy, Siddhartha Kamalakara

https://mlcollective.org/iclr-2021-open-collab-social/

Making AI research more inviting, inclusive, and accessible is a difficult task, but the movement to do so is close to many researchers' hearts. Progress toward democratizing AI research has been centered around making knowledge (e.g. class materials), established ideas (e.g. papers), and technologies (e.g. code) more accessible. However, open, online resources are only part of the equation. Growth as a researcher requires not only learning by consuming information individually, but hands-on practice whiteboarding, coding, plotting, debugging, and writing collaboratively, with either mentors or peers.

Of course, making ""collaborators"" more universally accessible is fundamentally more difficult than, say, ensuring all can access arXiv papers, because scaling people and research groups is much harder than scaling websites.

Can we nevertheless make access to collaboration itself more open? Can we flatten access to peers and mentors so the opportunities available to those at the best industrial and academic labs are more broadly available to all entrants to our burgeoning field? How to kick start remote, non-employment based research collaborations more easily? This social is designed to help you meet potential collaborators, find interesting ideas, and kick start your next project.

Daneil Chambers, Brian Liou

We will explore an offer negotiation webinar, where we will share what negotiation tactics to expect from companies, what mistakes we see candidates make when negotiating, how to understand your leverage in a negotiation, and more.Then, we will hear from Dr. Gael Varoquaux on making a career decision when he didn’t have a clue, on building his career at INRIA, and on effectively negotiating for what you need and deserve without hurting relationships. Gael will share stories as a top Researcher who now hires and manages hundreds of Ph.D.s and scientists.

Tianmin Shu, Xavier Puig

Social intelligence, the ability to understand other agents’ mental states, and interact with them, is at the core of human and machine intelligence. There has been an accelerating interest in machine social intelligence, or Social AI. However, as an emerging field, there has been little formal discussion on the challenges and advances in engineering machine social intelligence. This virtual gathering aims to provide an opportunity for people (both beginners and experts) to connect with one another, learn more about the advances in the field, and discuss scientific questions about the field. To this end, we have invited four amazing speakers, Natasha Jaques (Google Research, UC Berkeley), Dylan Hadfield-menell (UC Berkeley), Alex Trott (Salesforce Research), and Shari Liu (MIT), who will give exciting talks about social RL, value alignment, AI for economics, and the origins of human social intelligence. We will also host open discussion sessions that welcome different opinions. See our website for the schedule: https://social-intelligence-human-ai.github.io/ICLRSocial

Jennifer Hobbs, Sujoy Ganguly

“Lapsed” (aka. Former) Physicists are plentiful in the machine learning community. Inspired by Wine and Cheese seminars at many institutions, this BYOB event is an informal opportunity to connect with members of the community. Hear how others made the transition between fields, discuss how your physics training prepared you to switch fields, what synergies between physics and machine learning excite you the most, share your favorite physics jokes your computer science colleagues don’t get, and just meet other cool people. Open to everyone, not just physicists, you just have to tolerate our humor. Wine and Cheese encouraged, but not required.

William Agnew
Anil Goyal, Sudarsun Santhiappan, Jeshuren Chelladurai Ram Swaminathan

With the explosion of electronic data in the healthcare space, the industry needs to find the best ways to extract relevant information from the data to help various stakeholders, such as doctors, patients, hospitals, and insurance companies. In this social event, we would be discussing various challenges while dealing with real-world unstructured healthcare data and how BUDDI.AI uses ML, NLP tech to solve these challenges. Further, we would also discuss about how the extracted information is used in various use-cases like Medical Autocoding, Pharma covigilance, Clinical Decision Systems, Claim Denial Prediction, Clinical Trials, Clinical Documentation Improvement and Revenue Cycle Management.

Yuriy Nevmyvaka

Join prominent academics in an interactive round table discussion on a variety of top-of-mind topics in ML, including applications at financial institutions. Panelists will include

  • Professor Tomaso Aste from UCL
  • PhD candidate Christopher Jung from Penn
  • Professor Michael Kearns from Penn
  • Professor Aleksander Madry from MIT
  • Professor Ameet Talwalkar from CMU
  • Professor Irina Rish from MILA
  • Professor Stephen Roberts from Oxford

The conversation will be moderated by Morgan Stanley’s Head of ML Research Dr. Yuriy Nevmyvaka. Topics may include networks, big tech vs startup “rebel alliance”, DNN “primitives” for tabular/market data, fairness, privacy, explainability, NAS (for use cases beyond vision and NLP), ML Research careers in Finance, and more.

Sangeetha Marshathalli Siddegowda, Armina Stepan, Babak Ehteshami Bejnordi

We sit down virtually to discuss challenges and opportunities that arise from having to start a career in ML virtually, with tips and tricks for how to approach the application as well as the virtual onboarding into a new team.

We’ll discuss best practices for hiring and starting a career with special attention given to underrepresented groups in the AI field and the challenges/opportunities of remote work. During this presentation we'll hear from a recruiter that will give us tips and tricks on how to tell our professional story through an application. We'll also hear from two machine learning engineers, one which transitioned from a different industry and one who transitioned from academia.

Then we'll open the floor for discussions and networking.

Ronald Clark, Shuyu Lin

Roundtable Chatroom is designed to maximize the opportunity for ML researchers and practitioners to meet and chat with each other. This idea is inspired by the roundtable discussions (back in the physical conference era), where a table of participants engage with a discussion on a specific topic for around 10min and then move on to the next table for a discussion of another topic. By the end of the event, people have engaged in multiple interesting conversations and some valuable friendships will have begun.

We want to reproduce a similar experience online using Zoom’s breakout room feature to simulate different ‘roundtables’. To elevate the experience and ensure quality of discussion, we have also invited a list of mentors, who will host the discussion of various topics. The selected topics will cover a wide range of important issues to the ML community, including managing negative feedback, doing a start-up or not, what makes a paper go viral, how to maintain a good relationship with your supervisor etc.

Everybody is welcome to join but we do require sign-ups so we know how many people to expect. Please do so here by 5pm BST Thursday May 6: (https://forms.gle/NL46ZucM1JDmNnfk6)

Yiyuan Li, Rishabh Joshi

Language research is tightly connected with and widely benefits from the advance of learning approaches, namely language modeling, few-shot learning, meta-learning approaches and representation learning. ‘ML and Language’ is the place to exchange ideas and opinions for the frontier and future of those directions, and raise attention to the challenges. We want to bring researchers together to share ideas, foster collaborations and discuss their problems specifically in ML applied to language. Come and also join us in the session and https://bit.ly/3esHwVk.

Madhulika Srikumar, Rosie Campbell

There is an emerging consensus that AI researchers and scientists must consider the societal impact of their research. This idea is already taking hold with the introduction of Broader Impact Statements at NeurIPS last year and Paper Checklist Guidelines for authors at this year’s edition. This social will explore the practical challenges that arise for researchers with the introduction of these new requirements. The social will provide the ICLR community with the opportunity to engage with researchers who are thinking about these questions in an informal setting. Come armed with questions, find peers that are facing similar challenges in their research and learn more about resources for anticipating downstream consequences!

There will be two sessions to accommodate researchers from different time zones. Please go to gather.town.

Mementor Beta

The Mementor portal should help us scheduling spontaneous virtual mentor sessions during ICLR and beyond. Our goal is to enable mentorship opportunities for researchers in machine learning, both as mentors and mentees, with a special focus on under-represented minorities.

Our initial goal is to provide a platform to support conversations between mentors and mentees. The mode of operation initially will be a “lighter” version, where a mentor, at a time of their convenience, has a video call, which everybody willing as a mentee can join. No person-to-person commitment.

The mentorship session serves as a platform to share experiences. These could be technical and research related (e.g., research topics and technical discussions), or could be about scientific communication (e.g., paper writing, presentation, networking), or could also be mental health, burnouts, work ethics, PhD life etc. The goal is to facilitate sharing of experiences between members of the community which would not happen otherwise.

Note: The mentorship sessions are not a platform for self-promotion or promotion of products.