Anant Sahai
Professor
Studies information theory, wireless spectrum sharing and the theory of overparameterized machine learning.
Anant Sahai's lab at UC Berkeley works on learning theory & optimization and signal processing & information theory. 18 papers since 2023, 1 at top venues, mostly ISIT and NeurIPS. 19 alumni. Applications close .
At a glance
- papers, 2023–2026
- 5/ year
- at T1 venues
- 6%
- PhD students
- —
- MS students
- —
- alumni listed
- 19
Papers by yearfrom the lab homepage, 2023–2026
- 6 papers2026 · T1 0 · collecting
- 3 papers2025 · T1 0
- 6 papers2024 · T1 0
- 3 papers2023 · T1 1
T1 venuesOtherVenues: ISIT, NeurIPS
Research focuspaper titles that use each topic's words
- Learning theory & optimization3 papers
- overparameterized2 papers
- Signal processing & information theory0 papers
- sharing0 papers
- theory0 papers
Paper list
From the lab homepage.
20266 papers
- ICLR WorkshopPretraining with Masked Backstories in a Toy World (opens in a new tab)
- DySPANReproducibility in Spectrum Research: A Platform for Open Data, Shared Code, and Competitive Benchmarking (opens in a new tab)
- ICLR WorkshopSynthetic Error Injection Fails to Elicit Self-Correction In Language Models (opens in a new tab)
- IEEE Information Theory Workshop (ITW), acceptedThe Entropy of Floating-Point Numbers (opens in a new tab)
- IEEE Information Theory Workshop (ITW), acceptedThe Fourth Quadrant: A Stylized View of Benign Misfitting (opens in a new tab)
- IEEE Journal on Selected Areas in Information TheoryThe Thermodynamic Costs of Simple Linear Regression (opens in a new tab)
20253 papers
- HiLDDifferent simultaneous mechanisms for in-context recall have distinct learning dynamics (opens in a new tab)
- MOSSEmergence, pretraining loss and associative recall: a toy model (opens in a new tab)
- International Conference on Learning Representations (ICLR), AprProvable weak-to-strong generalization via benign overfitting (opens in a new tab)
20246 papers
- DySPANAutomatic Indoor-Outdoor Detection Using Signals of Opportunity (opens in a new tab)
- ACM Workshop on Hot Topics in NetworksFrom Foe to Friend: The Surprising Turn of Mega Constellations in Radio Astronomy (opens in a new tab)
- Mathematics of Operations ResearchOn the Impossibility of Convergence of Mixed Strategies with Optimal No-Regret Learning (opens in a new tab)
- ICML WorkshopPolynomial Regression as a Task for Understanding In-context Learning Through Finetuning and Alignment (opens in a new tab)
- DySPANTowards Data-Driven Policies in Spectrum Management (opens in a new tab)
- DySPANUsing Signals of Opportunity to Establish Trust in Distributed Spectrum Monitoring Systems (opens in a new tab)
20233 papers
- ACM Workshop on Hot Topics in NetworksAutomatic Calibration in Crowd-sourced Network of Spectrum Sensors (opens in a new tab)
- ISITLower Bounds for Multiclass Classification with Overparameterized Linear Models (opens in a new tab)
- NeurIPS (T1)Precise asymptotic generalization for multiclass classification with overparameterized linear models (opens in a new tab)
Where alumni went19 listed, 0% to industry
Industry 0Academia 3
Also to
- Middle East Technical University
- Tenure track at University of Wisconsin
- University of Toronto
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
- USNYU
Juan C. Perdomo
0 T1 - USCMU
Sivaraman Balakrishnan
- Learning theory & optimization
- USGeorgia Tech
Abhishek Shetty
- Learning theory & optimization
- SGNTU
Lipo Wang
- Data management & big data
- USMIT
Gregory Wornell
- Learning theory & optimization
- USUCLA
Guang Cheng
- Trustworthy & safe ML
Your notes
Save this lab to keep notes.
Data
Some numbers could not be verified:
- The member list could not be found on the lab site.
- Collected
- Verified
- Papers
- The lab homepage
- Members
- Not collected
- Alumni
- Lab homepage
Tiers are venue classifications, not lab rankings.
UC Berkeleycloses