Publications & Preprints

Anton Xue, Avishree Khare, Rajeev Alur, Surbhi Goel, Eric Wong (2024). Logicbreaks: A Framework for Understanding Subversion of Rule-based Inference.

Surbhi Goel, Abhishek Shetty, Konstantinos Stavropoulos, Arsen Vasilyan (2024). Tolerant Algorithms for Learning with Arbitrary Covariate Shift. NeurIPS 2024 [Spotlight].

Mahdi Sabbaghi, George Pappas, Hamed Hassani, Surbhi Goel (2024). Explicitly Encoding Structural Symmetry is Key to Length Generalization in Arithmetic Tasks.

Kan Xu, Hamsa Bastani, Surbhi Goel, Osbert Bastani (2024). Stochastic Bandits with ReLU Neural Networks. ICML 2024.

GuanWen Qiu, Da Kuang, Surbhi Goel (2024). Complexity Matters: Dynamics of Feature Learning in the Presence of Spurious Correlations. ICML 2024.

Benjamin L. Edelman, Ezra Edelman, Surbhi Goel, Eran Malach, Nikos Tsilivis (2024). The Evolution of Statistical Induction Heads: In-Context Learning Markov Chains. NeurIPS 2024.

Benjamin L. Edelman, Surbhi Goel, Sham Kakade, Eran Malach, Cyril Zhang (2023). Pareto Frontiers in Neural Feature Learning: Data, Compute, Width, and Luck. NeurIPS 2023 [Spotlight].

Surbhi Goel, Steve Hanneke, Shay Moran, Abhishek Shetty (2023). Adversarial Resilience in Sequential Prediction via Abstention. NeurIPS 2023.

Bingbin Liu, Jordan T. Ash, Surbhi Goel, Akshay Krishnamurthy, Cyril Zhang (2023). Exposing Attention Glitches with Flip-Flop Language Modeling. NeurIPS 2023 [Spotlight].

Sitan Chen, Zehao Dou, Surbhi Goel, Adam R. Klivans, Raghu Meka (2023). Learning Narrow One-Hidden-Layer ReLU Networks. COLT 2023.

Bingbin Liu, Jordan T. Ash, Surbhi Goel, Akshay Krishnamurthy, Cyril Zhang (2022). Transformers Learn Shortcuts to Automata. ICLR 2023 [notable-top-5%].

Surbhi Goel, Sham Kakade, Adam Tauman Kalai, Cyril Zhang (2022). Recurrent Convolutional Neural Networks Learn Succinct Learning Algorithms. NeurIPS 2022.

Boaz Barak, Benjamin L. Edelman, Surbhi Goel, Sham Kakade, Eran Malach, Cyril Zhang (2022). Hidden Progress in Deep Learning: SGD Learns Parities Near the Computational Limit. NeurIPS 2022.

Nikunj Saunshi, Jordan T. Ash, Surbhi Goel, Dipendra Misra, Cyril Zhang, Sanjeev Arora, Sham Kakade, Akshay Krishnamurthy (2022). Understanding Contrastive Learning Requires Incorporating Inductive Biases. ICML 2022.

Benjamin L. Edelman, Surbhi Goel, Sham Kakade, Cyril Zhang (2022). Inductive Biases and Variable Creation in Self-Attention Mechanisms. ICML 2022.

Jordan Ash, Cyril Zhang, Surbhi Goel, Akshay Krishnamurthy, Sham Kakade (2022). Anti-Concentrated Confidence Bonuses for Scalable Exploration. ICLR 2022.

Jordan Ash, Surbhi Goel, Akshay Krishnamurthy, Dipendra Misra (2022). Investigating the Role of Negatives in Contrastive Representation Learning. AISTATS 2022.

Jordan Ash, Surbhi Goel, Akshay Krishnamurthy, Sham Kakade (2021). Gone Fishing: Neural Active Learning with Fisher Embeddings. NeurIPS 2021.

Vardis Kandiros, Yuval Dagan, Nishanth Dikkala, Surbhi Goel, Constantinos Daskalakis (2021). Statistical Estimation from Dependent Data. ICML 2021.

Naman Agarwal, Surbhi Goel, Cyril Zhang (2021). Acceleration via Fractal Learning Rate Schedules. ICML 2021.

Surbhi Goel, Adam Klivans, Pasin Manurangsi, Daniel Reichman (2021). Tight Hardness Results for Training Depth-2 ReLU Networks. ITCS 2021.

Surbhi Goel, Aravind Gollakota, Adam Klivans (2020). Statistical-Query Lower Bounds via Functional Gradients. NeurIPS 2020.

Surbhi Goel, Adam Klivans, Frederic Koehler (2020). From Boltzmann Machines to Neural Networks and Back Again. NeurIPS 2020.

Surbhi Goel (2020). Learning Ising and Potts Models with Latent Variables. AISTATS 2020.

Surbhi Goel, Aravind Gollakota, Zhihan Jin, Sushrut Karmalkar, Adam Klivans (2020). Superpolynomial Lower Bounds for Learning One-Layer Neural Networks using Gradient Descent. ICML 2020.

Jessica Hoffmann, Soumya Basu, Surbhi Goel, Constantine Caramanis (2020). Learning Mixtures of Graphs from Epidemic Cascades. ICML 2020.

Omar Montasser, Surbhi Goel, Ilias Diakonikolas, Nathan Srebro (2020). Efficiently Learning Adversarially Robust Halfspaces with Noise. ICML 2020.

Ilias Diakonikolas, Surbhi Goel, Sushrut Karmalkar, Adam Klivans, Mahdi Soltanolkotabi (2020). Approximation Schemes for ReLU Regression. COLT 2020.

Surbhi Goel, Sushrut Karmalkar, Adam Klivans (2019). Time/Accuracy Tradeoffs for Learning a ReLU with respect to Gaussian Marginals. NeurIPS 2019 [Spotlight].

Surbhi Goel, Adam Klivans (2019). Learning Neural Networks with Two Nonlinear Layers in Polynomial Time. COLT 2019.

Surbhi Goel, Daniel Kane, Adam Klivans (2019). Learning Ising Models with Independent Failures. COLT 2019.

Surbhi Goel, Rina Panigrahy (2019). Recovering the Lowest Layer of Deep Networks with High Threshold Activations.

Matt Jordan, Naren Manoj, Surbhi Goel, Alex Dimakis (2019). Quantifying Perceptual Distortion of Adversarial Examples.

Surbhi Goel, Adam Klivans, Raghu Meka (2018). Learning One Convolutional Layer with Overlapping Patches. ICML 2018 [Long Talk].

Simon Du, Surbhi Goel (2018). Improved Learning of One-hidden-layer Convolutional Neural Networks with Overlaps.

Surbhi Goel, Adam Klivans (2017). Eigenvalue Decay Implies Polynomial-Time Learnability for Neural Networks. NeurIPS 2017.

Surbhi Goel, Varun Kanade, Adam Klivans, Justin Thaler (2017). Reliably Learning the ReLU in Polynomial Time. COLT 2017.