-->

About

I am currently an Applied Research Scientist at LinkedIn. Before, that I got my Computer Science PhD degree at University of California, Los Angeles, advised by professor Guy Van den Broeck. Before that I got my bachelors degree in Computer Science at UC Irvine speciallizing in both Algorithms and Intelligent Systems, accompanied by a Minor in Mathematics.

My research interests include:

  • Explainability and Fairness in AI & ML
  • Probabilistic Reasoning and Learning
  • Tractable Probabilistic Models
  • Machine Learning

Pasha Khosravi

Publications

[Pooladzandi2023]
Generating High Fidelity Synthetic Data via Coreset selection and Entropic Regularization.
Omead Pooladzandi, Pasha Khosravi, Erik Nijkamp, Baharan Mirzasoleiman.
In the Workshop on Synthetic Data for Empowering ML Research at NeurIPS, 2022.
Links
Paper

[KhosraviTPM22] Why Is This an Outlier? Explaining Outliers by Submodular Optimization of Marginal Distributions.
Pasha Khosravi, Antonio Vergari, Guy Van den Broeck.
In the 5th Workshop on Tractable Probabilistic Modeling at UAI (TPM 2022).
Links
Paper BibTeX

[NegoescuRecSys22] Sales Channel Optimization via Simulations based on Observational Data with Delayed Rewards: A Case Study at LinkedIn.
Diana Negoescu, Pasha Khosravi, Shadow Zhao, Nanyu Chen, Parvez Ahammad, Humberto Gonzalez.
In the Consequences+Reveal workshop at RecSys 2022.
Links
Workshop Page

[WangIJCAI21] Probabilistic Sufficient Explanations.
Eric Wang, Pasha Khosravi, Guy Van den Broeck.
In Proceedings of IJCAI, 2021.
Links
Paper Poster BibTeX

[DangAAAI21] Juice: A Julia Package for Logic and Probabilistic Circuits.
Meihua Dang, Pasha Khosravi, Yitao Liang, Antonio Vergari, Guy Van den Broeck.
In Proceedings of AAAI Conference (Demo Track), 2021.
Links
Paper BibTeX Github LogicCircuits.jl ProbabilisticCircuits.jl

[KhosraviArtemiss20] Handling Missing Data in Decision Trees: A Probabilistic Approach.
Pasha Khosravi, Antonio Vergari, YooJung Choi, Yitao Liang and Guy Van den Broeck.
In The Art of Learning with Missing Values Workshop at ICML (Artemiss), 2020.
Links
Paper BibTeX

[KhosraviNeurIPS19] On Tractable Computation of Expected Predictions.
Pasha Khosravi, YooJung Choi, Yitao Liang, Antonio Vergari and Guy Van den Broeck.
In Proceedings of NeurIPS, 2019.
Links
Paper BibTeX Poster

[KhosraviIJCAI19]
What to Expect of Classifiers? Reasoning about Logistic Regression with Missing Features.
Pasha Khosravi, Yitao Liang, YooJung Choi, Guy Van den Broeck.
In Proceedings of IJCAI, 2019.
Links
Paper BibTeX Poster Slides

Other/Preprints/Early Access
[RongIEEE2022]
Multi-SANA: Comparing measures of topological similarity for multiple network alignment.
Shiyue Rong, Weisheng Wang, Viet Ly, Pasha Khosravi, William C. Wu, Jing Chen, and Wayne B. Hayes
In IEEE Transactions on Evolutionary Computation, 2022.
Links
Paper BibTeX

[Wang2022]
On the current failure--but bright future--of topology-driven biological network alignment.
Siyue Wang, Xiaoyin Chen, Brent J Frederisy, Benedict A Mbakogu, Amy D Kanne, Pasha Khosravi, Wayne B Hayes
In Advances in Protein Chemistry and Structural Biology, 2022
Links
Paper BibTeX

[WangXXAI20] Towards Probabilistic Sufficient Explanations.
Eric Wang, Pasha Khosravi, Guy Van den Broeck.
In Extending Explainable AI Beyond Deep Models and Classifiers workshop at ICML (XXAI), 2020.
Links
Paper BibTeX

Experience

Industry
LinkedIn Software Engineer, Machine Learning, LinkedIn.
Focus on Root Cause Analysis (RCA), Measurement, and Experimentation Science. 		June 2023 - Now
LinkedIn Machine Learning Intern, Meta.
Probability Org. Ranking, Relevance, and Recommendations Team.


Sep 2022 - Dec 2022
LinkedIn Applied Research Data Science Intern, LinkedIn.
Data Science and Applied Research (DSAR) team. 		June 2022 - Sep 2022
LinkedIn Applied Research Data Science Intern, LinkedIn.
Data Science and Applied Research (DSAR) team. 		June 2021 - Sep 2021
Snap Research Intern, Snap Inc.
Data Insights and Governance 		June 2020 - Sep 2020
Facebook Software Engineering Intern, Facebook
Ads Delivery Infastructure 		June 2018 - Sep 2018
InfiniGlobe Sr. Developer, InfiniGlobe
Software development on LAMP Stack 		Oct 2017 - June 2018
AWS Software Engineering Intern, Amazon
Aws Lex 		June 2017 - Sep 2017
Kiersted Technical Consultant, Kiersted Systems
June 2015 - Sep 2016
Research
StarAI Graduate Student Researcher
Member of Statistical and Relational Artificial Intelligence Lab (StarAI).
Thesis: "Applications of Probabilistic Reasoning in Trustworthy AI: From Handling Missing Data to Explainability" 		Prof. Guy Van den Broeck
Sep 2018 - Now
UCI Undergraduate Research
Worked on explaining actions of Reinforcement Learning agents using black-box approaches such as LIME
Prof. Sameer Singh
Aug 2017 - June 2018
UCI Undergraduate Research
Worked on algorithm for finding best alignments between multiple biological networks such as Protein-to-Protein networks. 		Prof. Wayne Hayes
Jan 2016 - June 2018
Teaching
ICPC Coach ICPC Coach
Prepared material for the "ACM Class", the goal of the class was to improve students' problem solving, coding skills, and preparing them for competitions like ACM ICPC. Prepared short lectures and practice problems (twice a week meetings).
ACM @ UCI
Sep 2017 - June 2018
ICPC Coach Math Tutor
Tutored all lower division math classes including calculus, Linear Algebra, Differential Equations, and Discrete Math 		Irvine Valley College
Aug 2014 – May 2015

Projects

Juice

ProbabilisticCircuits.jl

Juice is a Julia package that offers researchers and practitioners efficient routines to construct, compile, learn and reason with Logic Circuits and Probabilistic Circuits. It provides a wide range of functionality that enable tractable logical and probabilistic reasoning.
NaCL

Naive Conformant Learning (NaCL)

Given a logistic regression (LR) model, NaCL learns a Naive Bayes model that conforms to the LR model while maximizing the joint feature likelihood on the dataset. The learned model gives us the ability to handle missing features, while getting benefits of higher accuracy of LR models.
Minecraft

Zombie Hunter

Used reinforcement learning and deep learning to build a agent to learn to fight zombies in Minecraft (Malmo platform)
Project in AI Class (CS 175, Spring 2017, UC Irvine)
Project Report Website
Thought by Professor Sameer Singh

Awards/Honors

ACM ICPC Ranked 56th in ACM ICPC World Finals, May 2017
Ranked 1st   in ACM ICPC Southern California Regional, Nov 2016
Ranked 3rd   in ACM ICPC Southern California Regional, Nov 2017
IEEExtreme Ranked 59th in IEEExtreme 10.0, Oct 2016 (1st in USA)
Ranked 24th in IEEExtreme 11.0, Oct 2017 (2nd in USA)
UCI Regent Regents' Scholar at UC Irvine.
"Most distinguished merit scholarship awarded by the University of California."