Sai Aparna Aketi
Computer Science · Purdue University West Lafayette
Publications
23
Citations
208
Est. group size
~1
Recurring co-author estimate
Active years
9
Publishing since 2018
Sai Aparna Aketi's research focuses on making distributed and decentralized machine learning more efficient and robust when data is spread unevenly across many devices or nodes (so-called 'non-IID' or heterogeneous data). Work spans techniques for reducing communication and computation costs during training, improving privacy (e.g., differentially private training for large language models), and handling unbalanced or changing data distributions in decentralized networks. Much of this research appears as preprints and conference/journal papers on algorithms for training neural networks across multiple machines without relying on a central server.
Publication output has grown from occasional papers before 2020 to a steadier pace of about 3 per year over the last five years, with a peak around 2023-2024.
Generated by claude-sonnet-5 from public bibliographic data · Jul 20, 2026
- Efficient DP-SGD for LLMs with Randomized Clipping
arXiv (Cornell University) · 2026
- Efficient DP-SGD for LLMs with Randomized Clipping
arXiv (Cornell University) · 2026
- SADDLe: Sharpness-Aware Decentralized Deep Learning with Heterogeneous Data
2025
- Cross-feature Contrastive Loss for Decentralized Deep Learning on Heterogeneous Data
2024
- Averaging Rate Scheduler for Decentralized Learning on Heterogeneous Data
arXiv (Cornell University) · 2024
- AdaGossip: Adaptive Consensus Step-size for Decentralized Deep Learning with Communication Compression
arXiv (Cornell University) · 2024
- SADDLe: Sharpness-Aware Decentralized Deep Learning with Heterogeneous Data
arXiv (Cornell University) · 2024
- Global Update Tracking: A Decentralized Learning Algorithm for Heterogeneous Data
arXiv (Cornell University) · 2023
- CoDeC: Communication-Efficient Decentralized Continual Learning
arXiv (Cornell University) · 2023
- Homogenizing Non-IID datasets via In-Distribution Knowledge Distillation for Decentralized Learning
arXiv (Cornell University) · 2023
- Cross-feature Contrastive Loss for Decentralized Deep Learning on Heterogeneous Data
arXiv (Cornell University) · 2023
- Global Update Tracking: A Decentralized Learning Algorithm for Heterogeneous Data
2023
- Low precision decentralized distributed training over IID and non-IID data
Neural Networks · 2022
- Neighborhood Gradient Clustering: An Efficient Decentralized Learning Method for Non-IID Data Distributions
arXiv (Cornell University) · 2022
- Low Precision Decentralized Distributed Training with Heterogeneous Data
arXiv (Cornell University) · 2021
- arXiv (Cornell University)×15
- Frontiers in Neuroscience×1
- IEEE Access×1
- Neural Networks×1
- IEEE Transactions on Cognitive and Developmental Systems×1
- Dong-Jun Han
Computer Science · Purdue University West Lafayette
- Rohit Parasnis
Computer Science · Purdue University West Lafayette
- Christopher G. Brinton
Computer Science · Purdue University West Lafayette
- Feijie Wu
Computer Science · Purdue University West Lafayette
- Wenzhi Fang
Computer Science · Purdue University West Lafayette
This profile was generated automatically from public scholarly data (OpenAlex). Group size and activity levels are estimates derived from co-authorship patterns.
Last updated Jul 20, 2026.
Claim or correct this profile