Alex Dimakis
Professor
Studies generative models, information theory and machine learning, including inverse problems with deep generative priors.
Alex Dimakis's lab at UC Berkeley works on signal processing & information theory and learning theory & optimization. 12 papers since 2023, 7 at top venues, mostly ICLR and ICCAD. 6 PhD students. 24 alumni, most often to Google and Amazon. Applications close .
At a glance
- papers, 2023–2026
- 3/ year
- at T1 venues
- 58%
- PhD students
- 6
- MS students
- 0
- alumni listed
- 24
Papers by yearfrom the lab homepage, 2023–2026
- 3 papers2026 · T1 3 · collecting
- 3 papers2025 · T1 3
- 0 papers2024 · T1 0
- 6 papers2023 · T1 1
T1 venuesOtherVenues: ICLR, ICCAD
Research focuspaper titles that use each topic's words
- Learning theory & optimization3 papers
- inverse problems2 papers
- Signal processing & information theory0 papers
- generative0 papers
- studies generative0 papers
Paper list
From the lab homepage.
20263 papers
- ICLR (T1)GEPA: Reflective Prompt Evolution Can Outperform Reinforcement Learning
- ICLR (T1)OpenThoughts: Data Recipes for Reasoning Models
- ICLR (T1)Terminal-Bench: Benchmarking Agents on Hard, Realistic Tasks in Command Line Interfaces
20253 papers
- ICLR (T1)Ambient Diffusion Posterior Sampling: Solving Inverse Problems with Diffusion Models Trained on Corrupted Data
- ICLR (T1)Infilling Score: A Pretraining Data Detection Algorithm for Large Language Models
- ICLR (T1)Language models scale reliably with over-training and on downstream tasks
20236 papers
- —Ambient Diffusion: Learning Clean Distributions from Corrupted Data
- Proc of Neural Information Processing Systems (NeurIPS), Dec Datasets and Benchmarks Track Selected for Oral presentation (ProjectDataComp: In search of the next generation of multimodal datasets
- Proceedings on the International Conference on Representation Learning (ICLRHotProtein: A Novel Framework for Protein Thermostability Prediction and Editing
- —Martingale Diffusion Models: Mitigating Sampling Drift by Learning to be Consistent
- ICCAD (T1)One-Dimensional Deep Image Prior for Curve Fitting of S-Parameters from Electromagnetic Solvers
- —Solving Inverse Problems Provably via Posterior Sampling with Latent Diffusion Models
Where alumni went24 listed, 68% to industry
Industry 13Academia 6
Most often to
- Amazon
- Google DeepMind
Also to
- MIT
- UC Berkeley
- Princeton University
Apply to UC BerkeleyFall 2027
UC Berkeley EECS Computer Science PhD
- Round
- Single round
- Opens
- —
- GRE
- not required
- English
- TOEFL/IELTS (non-English-country degrees)
- Fee
- See program page
Confirmed on the program page.
Similar labsnearest by research summary
- USCMU
Andrej Risteski
- Learning theory & optimization
- GBUCL
Arthur Gervais
- Trustworthy & safe ML
- USMIT
Ankur Moitra
- Trustworthy & safe ML
- USUT Austin
Adam Klivans
- Learning theory & optimization
- CAToronto
Murat Erdogdu
- Learning theory & optimization
- CNSJTU
Ziyang Zheng
- Generative & foundation models
Your notes
Save this lab to keep notes.
Data
Checked against the lab homepage.
- Collected
- Verified
- Papers
- The lab homepage
- Members
- Lab homepage
- Alumni
- Lab homepage
Tiers are venue classifications, not lab rankings.
UC Berkeleycloses