Jason D. Lee
Identifiers
- name variant Jason D. Lee 0.60 · backfill
Papers (39)
- Fine-Tuning Dynamics of In-Context Factual Recall in Transformers cs.LG · 2026 · author #3
- Sharp Capacity Thresholds in Linear Associative Memory: From Winner-Take-All to Listwise Retrieval stat.ML · 2026 · author #4
- The Power of Power Law: Asymmetry Enables Compositional Reasoning cs.AI · 2026 · author #3
- Sharp Capacity Scaling of Spectral Optimizers in Learning Associative Memory cs.LG · 2026 · author #5
- Discrepancies are Virtue: Weak-to-Strong Generalization through Lens of Intrinsic Dimension cs.LG · 2025 · author #4
- Medusa: Simple LLM Inference Acceleration Framework with Multiple Decoding Heads cs.LG · 2024 · author #5
- Lexicographic and Depth-Sensitive Margins in Homogeneous and Non-Homogeneous Deep Models stat.ML · 2019 · author #3
- Solving Non-Convex Non-Concave Min-Max Games Under Polyak-{\L}ojasiewicz Condition math.OC · 2018 · author #3
- Gradient Descent Finds Global Minima of Deep Neural Networks cs.LG · 2018 · author #2
- Provably Correct Automatic Subdifferentiation for Qualified Programs math.OC · 2018 · author #2
- Algorithmic Regularization in Learning Deep Homogeneous Models: Layers are Automatically Balanced cs.LG · 2018 · author #3
- Adding One Neuron Can Eliminate All Bad Local Minima stat.ML · 2018 · author #3
- Stochastic subgradient method converges on tame functions math.OC · 2018 · author #4
- Convergence of Gradient Descent on Separable Data stat.ML · 2018 · author #2
- On the Power of Over-parametrization in Neural Networks with Quadratic Activation cs.LG · 2018 · author #2
- Gradient Primal-Dual Algorithm Converges to Second-Order Stationary Solutions for Nonconvex Distributed Optimization math.OC · 2018 · author #2
- On the Convergence and Robustness of Training GANs with Regularized Optimal Transport cs.LG · 2018 · author #4
- Gradient Descent Learns One-hidden-layer CNN: Don't be Afraid of Spurious Local Minima cs.LG · 2017 · author #2
- Learning One-hidden-layer Neural Networks with Landscape Design cs.LG · 2017 · author #2
- First-order Methods Almost Always Avoid Saddle Points stat.ML · 2017 · author #1
- When is a Convolutional Filter Easy To Learn? cs.LG · 2017 · author #2
- An inexact subsampled proximal Newton-type method for large-scale machine learning cs.LG · 2017 · author #3
- Gradient Descent Can Take Exponential Time to Escape Saddle Points math.OC · 2017 · author #3
- Black-box Importance Sampling stat.ML · 2016 · author #2
- Sketching Meets Random Projection in the Dual: A Provable Recovery Algorithm for Big and High-dimensional Data cs.LG · 2016 · author #2
- Communication-Efficient Distributed Statistical Inference stat.ML · 2016 · author #2
- Matrix Completion has No Spurious Local Minimum cs.LG · 2016 · author #2
- Gradient Descent Converges to Minimizers stat.ML · 2016 · author #1
- A Kernelized Stein Discrepancy for Goodness-of-fit Tests and Model Evaluation stat.ML · 2016 · author #2
- Learning Halfspaces and Neural Networks with Random Initialization cs.LG · 2015 · author #2
- $\ell_1$-regularized Neural Networks are Improperly Learnable in Polynomial Time cs.LG · 2015 · author #2
- Distributed Stochastic Variance Reduced Gradient Methods and A Lower Bound for Communication Complexity math.OC · 2015 · author #1
- Selective Inference and Learning Mixed Graphical Models stat.ML · 2015 · author #1
- Communication-efficient sparse regression: a one-shot approach stat.ML · 2015 · author #1
- Scalable methods for nonnegative matrix factorizations of near-separable tall-and-skinny matrices cs.LG · 2014 · author #2
- Exact post-selection inference, with application to the lasso math.ST · 2013 · author #1
- On model selection consistency of regularized M-estimators math.ST · 2013 · author #1
- Proximal Newton-type methods for minimizing composite functions stat.ML · 2012 · author #1
- Learning Mixed Graphical Models stat.ML · 2012 · author #1
Mentions
- 1402.6964 #2 · backfill · confidence 0.70 Jason D. Lee
- 1311.6238 #1 · backfill · confidence 0.70 Jason D. Lee
- 2605.27774 #3 · arxiv_oai · confidence 0.70 Jason D. Lee
- 1305.7477 #1 · backfill · confidence 0.70 Jason D. Lee
- 1206.1623 #1 · backfill · confidence 0.70 Jason D. Lee
- 1205.5012 #1 · backfill · confidence 0.70 Jason D. Lee
Frequent Coauthors
- Michael I. Jordan 7 shared papers
- Simon S. Du 6 shared papers
- Yuekai Sun 5 shared papers
- Eshaan Nichani 3 shared papers
- Jonathan E. Taylor 3 shared papers
- Meisam Razaviyayn 3 shared papers
- Nathan Srebro 3 shared papers
- Qiang Liu 3 shared papers
- Rong Ge 3 shared papers
- Tengyu Ma 3 shared papers
- Aarti Singh 2 shared papers
- Barnabas Poczos 2 shared papers
- Benjamin Recht 2 shared papers
- Daniel Soudry 2 shared papers
- Juno Kim 2 shared papers
- Max Simchowitz 2 shared papers
- Maziar Sanjabi 2 shared papers
- Mor Shpigel Nacson 2 shared papers
- Sham Kakade 2 shared papers
- Suriya Gunasekar 2 shared papers