Akhil Kedia

14 papers A* 4A 1Journal 6Unranked 3
YearRankTypeTitle / Venue / Authors
2026 J jnl
LookaheadKV: Fast and Accurate KV Cache Eviction by Glimpsing into the Future without Generation.
CoRR
Jinwoo Ahn, Ingyu Seong, Akhil Kedia, Junhan Kim, Hyemi Jang, Kangwook Lee, Yongkweon Jeon
2025 — conf
Sparse Logit Sampling: Accelerating Knowledge Distillation in LLMs.
ACL (1)
Anshumann, Mohd Abbas Zaidi, Akhil Kedia, Jinwoo Ahn, Taehwak Kwon, Kangwook Lee, Haejun Lee, Joohyung Lee
2025 J jnl
Sparse Logit Sampling: Accelerating Knowledge Distillation in LLMs.
CoRR
Anshumann, Mohd Abbas Zaidi, Akhil Kedia, Jinwoo Ahn, Taehwak Kwon, Kangwook Lee, Haejun Lee, Joohyung Lee
2024 A* conf
Transformers Get Stable: An End-to-End Signal Propagation Theory for Language Models.
ICML
Akhil Kedia, Mohd Abbas Zaidi, Sushil Khyalia, JungHo Jung, Harshith Goka, Haejun Lee
2024 J jnl
Transformers Get Stable: An End-to-End Signal Propagation Theory for Language Models.
CoRR
Akhil Kedia, Mohd Abbas Zaidi, Sushil Khyalia, JungHo Jung, Harshith Goka, Haejun Lee
2024 J jnl
UnifiedCrawl: Aggregated Common Crawl for Affordable Adaptation of LLMs on Low-Resource Languages.
CoRR
Bethel Melesse Tessema, Akhil Kedia, Tae-Sun Chung
2022 A* conf
FiE: Building a Global Probability Space by Leveraging Early Fusion in Encoder for Open-Domain Question Answering.
EMNLP
Akhil Kedia, Mohd Abbas Zaidi, Haejun Lee
2022 J jnl
FiE: Building a Global Probability Space by Leveraging Early Fusion in Encoder for Open-Domain Question Answering.
CoRR
Akhil Kedia, Mohd Abbas Zaidi, Haejun Lee
2022 A* conf
You Only Need One Model for Open-domain Question Answering.
EMNLP
Haejun Lee, Akhil Kedia, Jongwon Lee, Ashwin Paranjape, Christopher D. Manning, Kyoung-Gu Woo
2021 — conf
Beyond Reptile: Meta-Learned Dot-Product Maximization between Gradients for Improved Single-Task Regularization.
EMNLP (Findings)
Akhil Kedia, Sai Chetan Chinthakindi, Wonho Ryu
2021 A conf
Keep Learning: Self-supervised Meta-learning for Learning from Inference.
EACL
Akhil Kedia, Sai Chetan Chinthakindi
2021 — conf
Learning to Generate Questions by Learning to Recover Answer-containing Sentences.
ACL/IJCNLP (Findings)
Seohyun Back, Akhil Kedia, Sai Chetan Chinthakindi, Haejun Lee, Jaegul Choo
2021 J jnl
You Only Need One Model for Open-domain Question Answering.
CoRR
Haejun Lee, Akhil Kedia, Jongwon Lee, Ashwin Paranjape, Christopher D. Manning, Kyoung-Gu Woo
2020 A* conf
NeurQuRI: Neural Question Requirement Inspector for Answerability Prediction in Machine Reading Comprehension.
ICLR
Seohyun Back, Sai Chetan Chinthakindi, Akhil Kedia, Haejun Lee, Jaegul Choo