Registration Check-in Desk closing at 6 pm. Badge pickup.
Solving the Right Problems: Making ML Models Relevant to Healthcare and the Life Sciences
The first-generation models for drug discovery and clinical applications were mostly direct modifications of algorithms developed for NLP, computer vision, and other well-established application areas. However, deployment of these models revealed the significant mismatch between their basic assumptions and the needs of these new life sciences applications. Examples include challenging generalization scenarios, unknown biases in the collected data, and the inability of domain experts to validate model predictions. In my talk, I will illustrate some of these problems, and introduce our initial solutions to them.
Un-Bookclub Haben: The Deafblind Woman Who Conquered Harvard Law
We’d love to come together for an un-bookclub at ICML. We’ve been learning a lot in the cross-continental book club out of the book Haben: The Deafblind Woman Who Conquered Harvard Law.
We’d love to give you the gift of connection, conversation, and reflection Haben gave us. We’ll invite Haben to stop by if she can (things are uncertain in the pandemic). We ask you to watch Haben's powerful talk at the National Book Festival in preparation: https://youtu.be/R57kh1ydSlk
Haben Girma, the author of the book, defines disability as an opportunity for innovation. She learned non-visual techniques for everything from dancing salsa to handling an electric saw. She developed a text-to-braille communication system that created an exciting new way to connect with people. Haben pioneered her way through obstacles, graduated from Harvard Law, and now uses her talents to advocate for disabled people.
Join us for a discussion on accessibility and intersectionality in the tech industry, and the roles and responsibilities of the machine learning community in building a culture where disabled people thrive."
Black in AI and Queer in AI Joint Social Event
A joint social with researching affinity groups Black in AI and Queer in AI celebrating the work of queer and black scientists. These events seek to fostercollaborations and mentorship among people from both communities, and discuss initiatives to increase the presence of black and queer people in the field of Artificial Intelligence.
There is a fast growing literature in econometrics on estimating causal effects in settings with panel or longitudinal data, building on the recent difference-in-differences and synthetic control literatures. This is driven by an empirical literature with many applications. These range from settings with few cross-sectional units and many or few time periods, to many cross-sectional units. Sometimes there are few, or even only one treated unit, and sometimes many. I will review some of this recent literature including some of the examples, focusing in particular on the synthetic difference in differences estimator, and some of the relations with matrix completion literature. I will also discuss implications for randomized experiments. I will discuss some of the remaining challenges in the literature.
Women in Artificial Intelligence and Machine Learning for Mental Health Applications
Artificial intelligence (AI) has strengthened its progress, and AI positions are the top emerging jobs worldwide; however, the talented women, who could help AI organizations achieve their ambitions, have been underrepresented in the field. The 2020 World Economic Forum report has shown that women account for only 26% of data and AI positions in the workforce. On the other hand, About two percent of the world's population suffers from various types of mental health disorders. Psychological health problems, e.g., depression, are among the ten principal causes of disability in all countries. Due to the large scale of the problem, tackling it falls under one of the 17 sustainable development goals of the United Nations. Thus, it is critical to foster research partnerships that result in the development of AI-driven solutions to measure, diagnose, and treat mental and neurodegenerative disorders.
To promote women's roles in AI for mental health application, we aim to organize a virtual social event for international women in AI and ML and a physical, social event for women in AI and ML in Montreal: Establish informal science relationships with women researchers in AI research centers around the world. Facilitate formal technology partnerships with women CEOs and women in AI companies for Mental Health around the world. Highlight impacts of research projects by women in AI and ML for Mental Health Applications Develop AI-based assessment tools in doing early detection and real-world objective measurement of mental health problems.
Organizing such events will allow women in AI and ML working for Mental Health applications to expand their knowledge regarding ongoing AI research projects and innovative AI applications. They can also receive valuable feedback from other AI communities worldwide.
Events Schedule We will endeavor to organize an online social event for ICML 2022. The events will generally include one or two long talks (20 or 30 minutes) with keynote speakers, three lightning talks (5 to 10 minutes), and a panel discussion. Diversity Commitment The events’ organizers share a commitment to diversity among the organizers and invited speakers in terms of gender, race, experience, institutional affiliation, and field. We also attempt to request diverse people from as many places as possible (e.g., mailing lists, social media) to promote the social event.
Interdisciplinary ML Mixer
In this in-person, 2-hour Social, we pair participants together based on differing types of experience in related subdisciplines of ML. These pairings would be determined by a confidential online form that participants would fill out upon entering the Social. For example, suppose Researcher A identifies as being highly experienced in Reinforcement Learning but has little to no experience in Optimization. Researcher A could then be paired with Researcher B, who has a great background in studying Optimization but has had no exposure to Reinforcement Learning. This cycle could repeat every 15 minutes so that every participant can meet a diverse set of researchers across a multitude of subdisciplines. The Social would end with a round table discussion talking about the various topics that they learned about and how it could influence their research moving forward.