Parnian Kassraie
Ph.D. Student, ETH Zurich
pkassraie [AT] ethz.ch
Bio
I am a 2nd year Ph.D. student at the department of Computer Science at ETH Zurich, fortunate to be advised by Andreas Krause and Peter Bühlmann. I am a member of ETH Institute for Machine Learning and associated with the ETH AI Center. I am interested in statistics, optimization, and theory of Machine Learning, often in the context of sequential decision-making and online learning.
I completed my M.Sc. at ETH, under the supervision of Andreas Krause and Fanny Yang, with a thesis on Contextual Neural Bandits. Before coming to ETH, I did a dual B.Sc. in Electrical Engineering and Computer Science at Sharif University of Technology. My thesis was on Visualizing Adversarial Attacks on Convolutional Neural Networks, which I wrote with Nassir Navab and Federico Tombari at TU Munich.
Publications
My work focuses on sample-efficient sequential decision-making in rich and structured domains. To address these problems, I adapt tools from deep learning [NN] and meta-learning/high-dimensional statistics [META].
Anytime Model Selection in Linear Bandits
Parnian Kassraie, Aldo Pacchiano, Nicolas Emmenegger, Andreas Krause
In Conference on Neural Information Processing Systems (NeurIPS), 2023
Graph Neural Network Powered Bayesian Optimization for Large Molecular Spaces
Miles Wang-Henderson*, Bartu Soyuer*, Parnian Kassraie, Andreas Krause, Ilija Bogunovic
In ICML Workshop on Structured Probabilistic Inference & Generative Modeling, 2023
Hallucinated Adversarial Control for Conservative Offline Policy Evaluation
Jonas Rothfuss*, Bhavya Sukhija*, Tobias Birchler*, Parnian Kassraie, Andreas Krause
In Conference on Uncertainty in Artificial Intelligence (UAI), 2023
Lifelong Bandit Optimization: No prior and No Regret
Felix Schur*, Parnian Kassraie*, Jonas Rothfuss, Andreas Krause
In Conference on Uncertainty in Artificial Intelligence (UAI), 2023
Instance-Dependent Generalization Bounds via Optimal Transport
Songyan Hou*, Parnian Kassraie*, Anastasis Kratsios*, Jonas Rothfuss*, Andreas Krause
arXiv Preprint, 2022
Graph Neural Network Bandits
Parnian Kassraie, Andreas Krause, Ilija Bogunovic
In Conference on Neural Information Processing Systems (NeurIPS), 2022
Meta-Learning Hypothesis Spaces for Sequential Decision-making
Parnian Kassraie, Jonas Rothfuss, Andreas Krause
In International Conference on Machine Learning (ICML), 2022
Also in Safe, Anytime-Valid Inference workshop (SAVI), 2022
Also in Adaptive Experimental Design and Active Learning Workshop (ReALML), 2022
Neural Contextual Bandits without Regret
Parnian Kassraie, Andreas Krause
In International Conference on Artificial Intelligence and Statistics (AISTATS), 2022
Anytime Model Selection in Linear Bandits
Parnian Kassraie, Aldo Pacchiano, Nicolas Emmenegger, Andreas Krause
In Conference on Neural Information Processing Systems (NeurIPS), 2023
Lifelong Bandit Optimization: No prior and No Regret
Felix Schur*, Parnian Kassraie*, Jonas Rothfuss, Andreas Krause
In Conference on Uncertainty in Artificial Intelligence (UAI), 2023
Meta-Learning Hypothesis Spaces for Sequential Decision-making
Parnian Kassraie, Jonas Rothfuss, Andreas Krause
In International Conference on Machine Learning (ICML), 2022
Also in Safe, Anytime-Valid Inference workshop (SAVI), 2022
Also in Adaptive Experimental Design and Active Learning Workshop (ReALML), 2022
Graph Neural Network Powered Bayesian Optimization for Large Molecular Spaces
Miles Wang-Henderson*, Bartu Soyuer*, Parnian Kassraie, Andreas Krause, Ilija Bogunovic
In ICML Workshop on Structured Probabilistic Inference & Generative Modeling, 2023
Hallucinated Adversarial Control for Conservative Offline Policy Evaluation
Jonas Rothfuss*, Bhavya Sukhija*, Tobias Birchler*, Parnian Kassraie, Andreas Krause
In Conference on Uncertainty in Artificial Intelligence (UAI), 2023
Instance-Dependent Generalization Bounds via Optimal Transport
Songyan Hou*, Parnian Kassraie*, Anastasis Kratsios*, Jonas Rothfuss*, Andreas Krause
arXiv Preprint, 2022
Graph Neural Network Bandits
Parnian Kassraie, Andreas Krause, Ilija Bogunovic
In Conference on Neural Information Processing Systems (NeurIPS), 2022
Neural Contextual Bandits without Regret
Parnian Kassraie, Andreas Krause
In International Conference on Artificial Intelligence and Statistics (AISTATS), 2022
News
- [18.06.23] Serving as the head teaching assistant for Probabilistic Artificial Intelligence.
- [07.09.23] Giving a talk at the Royal Statistical Society Conference 2023 on Model Selection for Sequential Inference and Decision-making.
- [19.06.23] Giving a talk at the PAC-Bayes Meets Interactive Learning Workshop on Anytime Model Selection in Linear Bandits.
- [01.06.23] Made it 2 years into the PhD.
- [08.05.23] LiBO and HAMBO to be presented at UAI 2023.
- [22.11.22] Attending NeurIPS@Paris. Presenting GNN Bandits.
- [03.11.22] New Preprint: Instance-Dependent Generalization Bounds via Optimal Transport.
- [12-16.09.22] Attending Quantifying Uncertainty Workshop @ Simons Institute. Presenting Meta-KeL and some new results.
- [17-23.07.22]
Attending ICML 2022, serving as a Session Chair. Presenting Meta-KeL at the main conference, and at ReALML.(visa arrived too late) - [21.06.22]
Invited to Deep Learning Theory Workshop @ Simons Institute.(visa arrived too late) - [01.06.22] Made it 1 year into the PhD.
- [30.05-03.06.22] Attending SAVI 2022. Presenting Meta-KeL.
- [28-30.03.22] Attending AISTATS 2022. Presenting Neural Contextual Bandits without Regret.
- [22.12.21] Giving a talk at ReACT 2021 on 'No-regret Bayesian Optimization with limited knowledge of the function space'.
- [01.06.21] Started a PhD.