Affinity Group Supported Pathways to ML Research Panel and Social
Ezinne Nwanko (UC Berkeley), Sanae Lofti (New York University), Maria Skoularidou (University of Cambridge), and Johan Obando Cerón (Mila, University of Montreal) will discuss their career paths and how their respective affinity groups helped them along their journey in a panel moderated by Pablo Samuel Castro (Google Brain). The panel is followed by an open social for AISTATS attendees hosted by Sarah Tan (Meta).
Climate simulations remain our best tools to predict global and regional climate change. Climate projection uncertainty stem, in part, from the poor or lacking representation of processes, such as turbulence, clouds that are not resolved on the grid of global climate models. The representation of these unresolved processes has been a bottleneck in improving climate projections. The explosion of climate data and the power of machine learning algorithms are suddenly offering new opportunities. For example, can data-driven machine learning methods help us deepen our understanding of these unresolved processes and simultaneously improve their representation in climate models to reduce climate projections uncertainty? In this talk, I will discuss the current state of climate modeling and its future, focusing on the advantages and challenges of using machine learning for climate projections. I will present some of our recent work in which we leverage tools from machine learning and deep learning to learn representations of unresolved ocean processes and improve climate simulations. Our work suggests that machine learning could open the door to discovering new physics from data and enhance climate predictions. Yet, many questions remain unanswered, making the next decade exciting and challenging for hybrid climate modeling.
Award Ceremony
Approximate Inference in Additive Factorial HMMs with Application to Energy Disaggregation
This paper considers additive factorial hidden Markov models, an extension to HMMs where the state factors into multiple independent chains, and the output is an additive function of all the hidden states. Although such models are very powerful, accurate inference is unfortunately difficult: exact inference is not computationally tractable, and existing approximate inference techniques are highly susceptible to local optima. In this paper we propose an alternative inference method for such models, which exploits their additive structure by 1) looking at the observed difference signal of the observation, 2) incorporating a “robust” mixture component that can account for unmodeled observations, and 3) constraining the posterior to allow at most one hidden state to change at a time. Combining these elements we develop a convex formulation of approximate inference that is computationally efficient, has no issues of local optima, and which performs much better than existing approaches in practice. The method is motivated by the problem of energy disaggregation, the task of taking a whole home electricity signal and decomposing it into its component appliances; applied to this task, our algorithm achieves state-of-the-art performance, and is able to separate many appliances almost perfectly using just the total aggregate signal.
WiML - CWS Social
Women in Machine Learning (WiML) and the Caucus for Women in Statistics (CWS) welcome AISTATS attendees to this social event. The event will start with icebreakers to encourage networking among participants. In the main part of the event there will be a Q&A with WiML sponsors, WiML board members and CWS board members in an open format: participants are encouraged to come with questions on various topics ranging from career advice or time-management to conducting research. The event is hosted by Tatjana Chavdarova (UC Berkeley), Christina Papadimitriou (Palo Alto Networks), and Jessica Kohlschmidt (Ohio State University Comprehensive Cancer Center).