3rd Women in Machine Learning Un-Workshop, ICML 2022

The 3rd WiML Un-Workshop is co-located with ICML on Monday, July 18th, 2022.

Speakers

Logistics

Breakout Sessions

Program

Call for Participation

Committee

FAQ

Code Of Conduct

Machine learning is one of the fastest growing areas of computer science research. Search engines, text mining, social media analytics, face recognition, DNA sequence analysis, speech and handwriting recognition, healthcare analytics are just some of the applications in which machine learning is routinely used.

In spite of the wide reach of machine learning and the variety of theory and applications, it covers, the percentage of female researchers is lower than in many other areas of computer science. Most women working in machine learning rarely get the chance to interact with other female researchers, making it easy to feel isolated and hard to find role models.

The annual Women in Machine Learning Workshop is the flagship event of Women in Machine Learning, primarily intended to foster active participant engagement in the program. This technical workshop gives female faculty, research scientists, and graduate students in the machine learning community an opportunity to meet, network and exchange ideas, participate in career-focused panel discussions with senior women in industry and academia and learn from each other. Underrepresented minorities and undergraduates interested in machine learning research are encouraged to attend. We welcome all genders; however, any formal presentations, i.e. talks and posters, are given by women. We strive to create an atmosphere in which participants feel comfortable to engage in technical and career-related conversations.

Now in its 3th year, the 2022 un-workshop is co-located with ICML.

Besides this un-workshop and annual workshop which is co-located with NeurIPS, Women in Machine Learning also organizes events such as lunch or social at the AISTATS or AAAI conferences, maintains a public directory of women active in ML, profiles the research of women in ML, and maintains a list of resources for women working in ML.

Invited Speakers

Emma Brunskill

Emma Brunskill is an associate professor in the Computer Science Department at Stanford University. Her goal is to create AI systems that learn from few samples to robustly make good decisions, motivated by our applications to healthcare and education. Her work has been honored by early faculty career awards (National Science Foundation, Office of Naval Research, Microsoft Research) received several best research paper nominations (CHI, EDMx3) and awards (UAI, RLDM, ITS).

Celestine Mendler-Dünner

Celestine Mendler-Dünner is a research group lead at the Max Planck Institute for Intelligent Systems in Tübingen. Her research focuses on the role of society in the study of computation, taking into account actions and reactions of individuals when analyzing and designing algorithmic systems. Prior to joining MPI-IS Celestine was a SNSF postdoctoral fellow at UC Berkeley, and a predoctoral researcher at IBM Research Zurich. She obtained her PhD from ETH Zurich where she was awarded the ETH medal and the Fritz Kutter prize for the academic as well as the industrial impact of her research.

Yixin Wang

Yixin Wang is an LSA Collegiate Fellow and an assistant professor of statistics at the University of Michigan. She works in the fields of Bayesian statistics, machine learning, and causal inference. Her research has received several awards, including the INFORMS data mining best paper award, Blackwell-Rosenbluth Award from the junior section of ISBA, student paper awards from ASA Biometrics Section and Bayesian Statistics Section, and the ICSA conference young researcher award.

Location

This workshop will be hybrid, co-located with ICML at the Baltimore Convention Center, Baltimore, Maryland USA.

Type of registration required to attend

Any type of in-person registration (tutorial / workshop / conference / all) grants you in-person access to the un-workshop. Also, an in-person registration includes access to the virtual one.

Breakout Sessions

During the day of the WiML Un-Workshop @ ICML 2022 there will be three different Breakout Sessions. We list the sessions, topics, and leaders.

BreakoutGhoshehBreakout Session #1 (9.10AM - 10.10AM)

IN-PERSON Breakout Sessions
- - Machine learning real-time applications in health. Leader: Dania Humaidan, Co-leader: Cansu Sen.
VIRTUAL Breakout Sessions
- - Deep Generative Models for Electronic Health Records. Leader: Ghadeer Ghosheh, Co-leader: Tingting Zhu.
  - Affective Computing: A Computational Perspective. Leader: Shreya Ghosh, Co-lead: Garima Sharma.
  - Introducing geometry awareness in deep networks. Leader: Ankita Shukla.

Breakout Session #2 (11.05AM - 12.05AM)

IN-PERSON Breakout Sessions
- - Challenges and opportunities in certified auditing of ML models. Leader: Chhavi Yadav.
  - Robustness of Deep Learning Models to Distribution Shift. Leader: Polina Kirichenko, Co-leads: Shiori Sagawa, Sanae Lofti.
VIRTUAL Breakout Sessions
- - Knowledge Distillation through the lense of the capacity gap problem. Leader: Ibtihel Amara, Co-lead: Samrudhdhi Rangrej, Zahra Vaseqi.
  - Improving AI Education. Leader: Mary Smart, Co-lead: Stefania Druga.
  - Statistical Inference & Applications to Machine Learning. Leader: Lilian Wong, Co-lead: Po-ling Loh.

Breakout Session #3 (15.25 - 16.25)

IN-PERSON Breakout Sessions
- - Robustness of Machine Learning. Leader: Yao Qin
  - Towards efficient and robust deep learning training. Leader: Wenhan Xia.
VIRTUAL Breakout Sessions
- - Machine Learning for Physical Sciences. Leader: Taoli Cheng.
  - Limitations of explainable/interpretable AI: frontiers and boundaries for future advancement. Leader: Haoyu Du, Co-lead: Peiyuan Zhou, Annie Lee, Rainah Khan.
  - Detection of Unseen Classes of different Domains using Computer Vision. Leader: Asra Aslam.

The program follows the following color scheme: talks, breakout sessions, poster sessions, mentoring sessions, program break, sponsor talks, panel discussion. All invited talk titles, and invited speaker/mentor/panelist names are *clickable*. The majority of the program will be streamed and occur synchronously in-person and virtually, except if marked as in-person/virtual only.

You can find the zoom links and livestream on the WiML workshop page of the ICML website.

08:30 Introduction & Opening Remarks, Vinitra Swamy

all-day Virtual Sponsor Booths, [DeepMind, D.E. Shaw Research, Home Depot, Microsoft Research]

all-day In-Person Sponsor Booths, [DeepMind, Google, QuantumBlack]

08:45 Desiderata for Representation Learning: A Causal Perspective, Yixin Wang [Invited Talk]

Abstract: Representation learning constructs low-dimensional representations to summarize essential features of high-dimensional data like images and texts. Ideally, such a representation should efficiently capture non-spurious features of the data. It shall also be disentangled so that we can interpret what feature each of its dimensions capture. However, these desiderata are often intuitively defined and challenging to quantify or enforce.
In this talk, we take on a causal perspective of representation learning. We show how desiderata of representation learning can be formalized using counterfactual notions, enabling metrics and algorithms that target efficient, non-spurious, and disentangled representations of data. We discuss the theoretical underpinnings of the algorithm and illustrate its empirical performance in both supervised and unsupervised representation learning.
Joint work with Michael Jordan.

09:10 Breakout session

[in-person only] Machine learning real-time applications in health (Leaders: Dania Humaidan, Cansu Sen)
[hybrid] Introducing geometry awareness in deep networks (Leader: Ankita Shukla)
[hybrid] Affective Computing: A Computational Perspective (Leaders: Shreya Ghosh, Garima Sharma)
[hybrid] Deep Generative Models for Electronic Health Records (Leaders: Ghadeer Ghosheh)

10:10 Poster Session

10:40 Emma Brunskill [Invited Talk]

11:05 Breakout session

[in-person only] Challenges and opportunities in certified auditing of ML models (Leader: Chhavi Yadav)
[in-person only] Robustness of Deep Learning Models to Distribution Shift (Leaders: Polina Kirichenko, Shiori Sagawa)
[hybrid] Knowledge Distillation through the Lens of the Capacity Gap Problem (Leaders: Ibtihel Amara, Samrudhdhi Rangrej, Zahra Vaseqi)
[hybrid] Improving AI Education (Leaders: Mary Smart, Stefania Druga)
[hybrid] Statistical Inference & Applications to Machine Learning (Leaders: Lilian Wong, Po-ling Loh)

12:05 Mentoring Roundtables [in-person only] /// Mentoring Panel [virtual only]

Table 1: Choosing between academia and industry Amy Zhang & Lauren Gardiner
Mentors: Jigyasa Grover, Ciara Pike-Burke, Nika Haghtalab, Po-Ling Loh, Hermina Petric Maretic

Table 2: Finding mentors and taking on mentorship roles throughout your career / Celestine Mendler-Dünner & Cyril Zhang
Moderator: Sinead Williamson

Table 3: Establishing and maintaining collaborations Surbhi Goel & Max Simchowitz

Table 4: Work-life Balance Ioana Bica & Kishore Kumar

13:05 Lunch Break, joint with NewInML [in-person only] /// Virtual Sponsor Booths [virtual only]

14:40 Harnessing the power of Hybrid Intelligence, Maria Olivia Lihn [QuantumBlack Sponsor Talk]

14:55 Building embodied agents that can learn from their environments and humans, Kavya Srinet [Meta Platforms Sponsor Talk]

15:10 Machine Learning at Apple, Tatiana Likhomanenko [Apple Sponsor Talk]

15:25 Breakout session

[in-person only] Robustness of Machine Learning (Leader: Yao Qin)
[in-person only] Distributionally robust Reinforcement Learning (Leaders: Laixi Shi, Mengdi Xu)
[hybrid] Machine Learning for Physical Sciences (Leader: Taoli Cheng)
[hybrid] Limitations of explainable/interpretable AI: frontiers and boundaries for future advancement (Leaders: Haoyu Du, Peiyuan Zhou, Annie Lee, Rainah Khan)
[hybrid] Detection of Unseen Classes of different Domains using Computer Vision (Leader: Asra Aslam)

16:30 Poster Session, joint with LXAI

17:00 Social dynamics in prediction, Celestine Mendler-Dünner [Invited Talk]

Abstract: Algorithmic predictions inform consequential decisions, incentivize strategic actions, and motivate precautionary measures. As such, predictions used in societal systems not only describe the world they aim to predict, but they have the power to change it; a prevalent phenomenon often neglected in theories and practices of machine learning. In this talk, I will introduce a risk minimization framework, called performative prediction, that conceptualizes this phenomenon by allowing the predictive model to influence the distribution over future data. This problem formulation elucidates different algorithmic solution concepts, optimization challenges, and offers a new perspective on prediction. In particular, I will discuss how performative prediction allows us to articulate the difference between learning from a population and steering a population through predictions, facilitating an emerging discourse on the topic of power of predictive systems in digital economies.

17:25 Best Practices for Research: Increasing Efficiency and Research Impact, and Navigating Hybrid Collaborations [Panel]

Panelists: Amy Zhang, Surbhi Goel, Agni Kumar
Moderator: Ioana Bica

18:25 Closing Remarks, Tatjana Chavdarova

Note: Please navigate the 'Program' menu in the slidebar at the top to find more details about speakers, panelist and mentors.

Surbhi Goel

Surbhi Goel is currently a postdoctoral researcher at Microsoft Research NYC. In Spring 2023, she will be starting as the Magerman Term Assistant Professor of Computer and Information Science at University of Pennsylvania. Prior to this, she received her Ph.D. from the Department of Computer Science at the University of Texas at Austin where she was advised by Adam Klivans. Her work lies at the intersection of machine learning and theoretical computer science, with a focus on developing the statistical and computational foundations of modern machine learning paradigms.

Among other honors, she is a recipient of UT Austin's Bert Kay Dissertation award, a J.P. Morgan AI PhD fellowship, and a Simons-Berkeley research fellowship. She has been recognized as a Rising Star in ML by University of Maryland and in EECS by UIUC. She is actively involved in service and outreach through her role as the co-founder of Learning Theory Alliance (LeT-All), a community building and mentorship initiative for the learning theory community.

Amy Zhang

Amy is a postdoctoral scholar at UC Berkeley and a research scientist at Facebook AI Research, and is starting as an assistant professor at UT Austin in the ECE department in Spring 2023. She works on state abstractions, model-based reinforcement learning, representation learning, and generalization in RL. She did her PhD at McGill University and Mila - Quebec AI Institute, co-supervised by Joelle Pineau and Doina Precup. She also has an M.Eng. in EECS and dual B.Sci. degrees in Mathematics and EECS from MIT.

Agni Kumar

Agni is an Applied Research Scientist on Apple’s Health AI team. She studied at MIT, graduating with an M.Eng. in Machine Learning and B.S. degrees in Mathematics and Computer Science. Her thesis on modeling the spread of healthcare-associated infections led to joining projects at Apple with applied health focuses, specifically on understanding cognitive decline from device usage data and discerning respiratory rate from wearable microphone audio. She has published hierarchical reinforcement learning research and predictive modeling work in conferences and journals, including CHIL, EMBC, PLOS Computational Biology, and Telehealth and Medicine Today. She was a workshop organizer for ICML’s first “Computational Approaches to Mental Health” workshop in 2021. She has also volunteered at WiML workshops and served as a reviewer for NeurIPS. For joy, Agni leads an Apple-wide global diversity network about encouraging mindfulness to find peace each day.

Ioana Bica (Moderator)

Ioana is a rising fifth-year PhD student at the University of Oxford and at the Alan Turing Institute, advised by Prof. Mihaela van der Schaar. Her PhD research focuses on building machine learning methods for improving and understanding decision making. To achieve this, she have worked on developing causal inference methods capable of estimating the individualized effect of interventions (e.g. actions or treatments) from observational data.

Her research experience also includes an internship at DeepMind where she has been working with Jovana Mitrović on self-supervised learning and causality with the aim of learning better representations for objects in images.

Prior to her PhD, she completed a Bachelor’s degree and a Master’s degree in Computer Science at the University of Cambridge where she worked with Prof. Pietro Liò on multi-modal data integration and unsupervised learning for genomics data. During this time, she has also interned at Google four times.

Amy Zhang

Celestine Mendler-Dünner

Celestine is a research group lead at the Max Planck Institute for Intelligent Systems in Tübingen. Her research focuses on the role of society in the study of computation, taking into account actions and reactions of individuals when analyzing and designing algorithmic systems.

Surbhi Goel

Surbhi is currently a postdoctoral researcher at Microsoft Research NYC. In Spring 2023, she will be starting as the Magerman Term Assistant Professor of Computer and Information Science at University of Pennsylvania. Prior to this, she received her Ph.D. from the Department of Computer Science at the University of Texas at Austin where she was advised by Adam Klivans. Her work lies at the intersection of machine learning and theoretical computer science, with a focus on developing the statistical and computational foundations of modern machine learning paradigms.

Ioana Bica

Ioana is a rising fifth-year PhD student at the University of Oxford and at the Alan Turing Institute, advised by Prof. Mihaela van der Schaar. Her research focuses on building machine learning methods for improving and understanding decision making.

Lauren Gardiner

Lauren is Senior Applied Research Scientist in the Health AI team at Apple.

Cyril Zhang

Cyril is a Senior Researcher at Microsoft Research NYC. His research interests include sequential prediction and decision-making, the theory and practice of optimization (especially in deep learning), and the synthesis of these topics (especially in language models).

Max Simchowitz

Max is a postdoc in Russ Tedrake's group at MIT. His recent work has focused on the theoretical foundations of online control and reinforcement learning, with past research ranging broadly across topics in adaptive sampling, multi-arm bandits, complexity of convex and non-convex optimization, and fairness in machine learning. He is currently interested in developing rigorous, theoretical guarantees for nonlinear control, wherever possible.

Kishore Kumar

Kumar is a Data Science and Analytics Lead at Amazon Prime video. He strives to solve complex business problems using advanced Machine Learning Algorithms, and has 10+ years of overall experience across multiple sectors.