8 papers · 1 filter
When Hard Negatives Hurt: Bridging the Generative-Discriminative Gap in Hard Negative Synthesis for Retrieval
Zhicheng Zhang, Jiwei Tang, Kuicai Dong +9
Hard negative mining has become the dominant strategy for training retrievers, yet it faces intrinsic limitations: negatives are bounded by corpus availability, selected by retriev…
Beyond Position Bias: Shifting Context Compression from Position-Driven to Semantic-Driven
Jiwei Tang, Zhijing Huang, Xinyu Zhang +5
Large Language Models (LLMs) have demonstrated exceptional performance across diverse tasks. However, their deployment in long-context scenarios faces high computational overhead a…
PoC: Performance-oriented Context Compression for Large Language Models via Performance Prediction
Runsong Zhao, Shilei Liu, Jiwei Tang +8
While context compression can mitigate the growing inference costs of Large Language Models (LLMs) by shortening contexts, existing methods that specify a target compression ratio…
COMI: Coarse-to-fine Context Compression via Marginal Information Gain
Jiwei Tang, Shilei Liu, Zhicheng Zhang +4
Large Language Models (LLMs) have demonstrated exceptional capabilities across diverse tasks. However, their deployment in long context scenarios remains hindered by computational…
Read As Human: Compressing Context via Parallelizable Close Reading and Skimming
Jiwei Tang, Shilei Liu, Zhicheng Zhang +9
Large Language Models (LLMs) demonstrate exceptional capability across diverse tasks. However, their deployment in long-context scenarios is hindered by two challenges: computation…
Data Distribution Matters: A Data-Centric Perspective on Context Compression for Large Language Model
Kangtao Lv, Jiwei Tang, Langming Liu +7
The deployment of Large Language Models (LLMs) in long-context scenarios is hindered by computational inefficiency and significant information redundancy. Although recent advanceme…
CoMeT: Collaborative Memory Transformer for Efficient Long Context Modeling
Runsong Zhao, Shilei Liu, Jiwei Tang +8
The quadratic complexity and indefinitely growing key-value (KV) cache of standard Transformers pose a major barrier to long-context processing. To overcome this, we introduce the…
Length-Adaptive Interest Network for Balancing Long and Short Sequence Modeling in CTR Prediction
Zhicheng Zhang, Zhaocheng Du, Jieming Zhu +8
User behavior sequences in modern recommendation systems exhibit significant length heterogeneity, ranging from sparse short-term interactions to rich long-term histories. While lo…