Aditya Krishna Menon

Research Scientist@Google·美国·OpenReview

研究方向

Long-tail learning · class imbalance · Distillation · label smoothing · Deep retrieval · Proper losses · surrogate losses · Cascades · routing · speculative decoding

6.5

合作者 (20)

Harikrishna Narasimhan

Aditya Krishna Menon

Universal Model Routing for Efficient LLM Inference

A Little Help Goes a Long Way: Efficient LLM Training by Leveraging Small LMs

Better autoregressive regression with LLMs via regression-aware fine-tuning

Bipartite Ranking From Multiple Labels: On Loss Versus Label Aggregation

Faster Cascades via Speculative Decoding

A Little Help Goes a Long Way: Efficient LLM Training by Leveraging Small LMs