LabCompass

Kaiyi Ji

Computer Science · The Ohio State University

OpenAlex ORCID
Early career · publishing since 2017
Stochastic Gradient Optimization TechniquesSparse and Compressive Sensing TechniquesDomain Adaptation and Few-Shot LearningAdvanced Bandit Algorithms ResearchMachine Learning and ELM

Publications

97

Citations

648

Est. group size

—

Recurring co-author estimate

Active years

10

Publishing since 2017

Research summary
AI-generated

Kaiyi Ji works on the mathematical foundations and algorithms behind modern machine learning, including bilevel optimization (problems with nested learning objectives), multi-task learning, continual learning (systems that learn new tasks without forgetting old ones), and federated learning (training across distributed devices). Much of the work is theoretical, analyzing convergence guarantees and complexity bounds for optimization methods used to train large models, alongside applied projects in areas like robotics and multi-task ranking systems.

Bilevel and stochastic optimization theoryContinual and lifelong learningMulti-task learning and optimizationFederated learningMeta-learning and few-shot adaptation

Publication output has fluctuated over the past decade with peaks around 2018, 2020, and 2023, but has remained fairly steady in recent years, averaging about 10 publications annually over the last five years.

Generated by claude-sonnet-5 from public bibliographic data · Jul 20, 2026

Publication cadence
Publications per year over the last 10 years — averaging 9.8/year recently
2017: 1 publication172018: 13 publications182019: 9 publications192020: 17 publications17202021: 8 publications212022: 6 publications222023: 14 publications232024: 12 publications242025: 10 publications252026: 7 publications26
Recent publications
Publishes in
  • arXiv (Cornell University)×56
  • Neural Information Processing Systems×2
  • IEEE Transactions on Information Theory×2
  • Desalination and Water Treatment×2
  • ACM SIGMETRICS Performance Evaluation Review×2
Similar researchers
By research-topic overlap

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 19, 2026.

Claim or correct this profile