I mostly run up hills but occasionally I do other things; here is a list of some worth mentioning


Seeker: Real Time Interactive Search, 2019

Abstract: This paper introduces Seeker, a system that allows users to adap- tively refine search rankings in real time, through a series of feed- backs in the form of likes and dislikes. When searching online, users may not know how to accurately describe their product of choice in words. An alternative approach is to search an embedding space, allowing the user to query using a representation of the item (like a tune for a song, or a picture for an object). However, this approach requires the user to possess an example representation of their de- sired item. Additionally, most current search systems do not allow the user to dynamically adapt the results with further feedback. On the other hand, users often have a mental picture of the desired item and are able to answer ordinal questions of the form: “Is this item similar to what you have in mind?” With this assumption, our algorithm allows for users to provide sequential feedback on search results to adapt the search feed. We show that our proposed approach works well both qualitatively and quantitatively. Unlike most previous representation-based search systems, we can quan- tify the quality of our algorithm by evaluating humans-in-the-loop experiments. Link

Authors: Ari Biswas, Thai T Pham, Michael Vogelsong, Benjamin Snyder, Houssam Nassif

Adverserial PCA, 2017

Abstract: This paper studies the following question: where should an adversary place an outlier of a given magnitude in order to maximize the error of the subspace estimated by PCA? We give the exact location of this worst possible outlier, and the exact expression of the maximum possible error. Equivalently, we determine the information-theoretic bounds on how much an outlier can tilt a subspace in its direction. This in turn provides universal (worst-case) error bounds for PCA under arbitrary noisy settings. Our results also have several implications on adaptive PCA, online PCA, and rank-one updates. We illustrate our results with a subspace tracking experiment.Link

Authors: Daniel Pimentel-Alcaron, Ari Biswas, Claudia R Solis-Lemus

Hands on Introduction to Computer Science at the Freshman level, 2014.

Abstract: This paper details the creation of a hands-on introduction course that reflects the dramatic growth and diversity in computer science. Our aim was to enable students to get an end-to-end perspective on computer system design by building one. We report on a two-year exercise in using the Arduino platform to build a series of hands-on projects. We have used these projects in two course instances, and have obtained detailed student feedback, which we analyze and present in this paper. The instructions, code and videos developed are available open-source. Link

Authors: Raghuman Balasubramanian, Zachary York, Matthew Doran, Ari Biswas, Timur Girgin and Professor Karu Sankaralingam



Amazon Shop By look

Lead Machine Learning scientist Shop by look produced by amazon.


Former contributer. Next is a real time distributed system for launching and analysing active learning algorithms. Link

Durian, 2016

In our quest to understand how people behave and think and to somehow quantify this understanding, we come upon the problem of modeling the relationships between objects within a person’s mind. For example, there are many aspects to how peoplethink about different fruits, but one popular comic has suggested a 2-dimensional model in which one axis represents the difficulty of preparation and the other represents tastiness (where bananas might end up at the top right, being both very easy and very tasty, while durian might be at the opposite corner). It is an interesting problem to try to experimentally learn such a model (with or without prescribed axes) for any given person or population. We present Durian, an experimental framework to allow a scientist to directly query subjects’ 2D mental models of a collection of objects using a variety of algorithms to perform the queries and to generalise the model to further data.

Link Needs clean up (code to be uploaded soon)


Set 1-6 complete. Cleanup and 7 pending. Code to be posted soon