History-aware cross-attention rewards for multi-turn and chain-of-thought fine-tuning claim +2 to +4 percent task gains and 3x latency wins, but the CoT reward as written is not computable for T5.
Of Spiky SVDs and Music Recommendation
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
abstract
The truncated singular value decomposition is a widely used methodology in music recommendation for direct similar-item retrieval or embedding musical items for downstream tasks. This paper investigates a curious effect that we show naturally occurring on many recommendation datasets: spiking formations in the embedding space. We first propose a metric to quantify this spiking organization's strength, then mathematically prove its origin tied to underlying communities of items of varying internal popularity. With this new-found theoretical understanding, we finally open the topic with an industrial use case of estimating how music embeddings' top-k similar items will change over time under the addition of data.
citation-role summary
background 1
citation-polarity summary
fields
cs.CL 1years
2025 1verdicts
REJECT 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
History-Aware Cross-Attention Reinforcement: Self-Supervised Multi Turn and Chain-of-Thought Fine-Tuning with vLLM
History-aware cross-attention rewards for multi-turn and chain-of-thought fine-tuning claim +2 to +4 percent task gains and 3x latency wins, but the CoT reward as written is not computable for T5.