Recommendation & Information Retrieval - Weekly Roundup

Multi-head attention applied to network depth unlocks substantial performance gains and compute efficiency in large MoE models.

Irrelevant text distractors cripple mixed-modality retrieval performance more than images, revealing a text modality preference that degrades performance in a V-shaped curve.

Standard two-level softmax sampling is demonstrably biased; corrected methods achieve near-exact softmax approximations with negligible overhead.

Adversarial shifts in bandit arm means, not just gaps, break standard best-arm identification algorithms, requiring a novel importance-weighting approach for robust performance.

Stored patch vectors from visual document retrieval systems can be inverted to reconstruct original pages, revealing sensitive information and enabling source page identification with high accuracy.

Multi-head attention applied to network depth unlocks substantial performance gains and compute efficiency in large MoE models.

Irrelevant text distractors cripple mixed-modality retrieval performance more than images, revealing a text modality preference that degrades performance in a V-shaped curve.

Standard two-level softmax sampling is demonstrably biased; corrected methods achieve near-exact softmax approximations with negligible overhead.

Adversarial shifts in bandit arm means, not just gaps, break standard best-arm identification algorithms, requiring a novel importance-weighting approach for robust performance.

Stored patch vectors from visual document retrieval systems can be inverted to reconstruct original pages, revealing sensitive information and enabling source page identification with high accuracy.

Generative recommender training strategies that append missed targets can actively harm ranking performance by introducing detrimental probability competition.

Black-box MIPS oracles can approximate attention with $O(1/\varepsilon^2)$ keys by augmenting queries and keys, bypassing $\sqrt{n}$ and $\log n$ dependencies.

Intermediate retrieval signals, not just final outcomes, unlock efficient credit assignment for training LLM search agents.

AI agents can now build and reuse persistent, artifact-specific knowledge, dramatically improving their ability to find relevant information for complex tasks and reducing operational costs.

LLMs can learn to actively compute evidence by routing through retrieval and computation, dramatically improving performance on complex, multi-source reasoning tasks.

A single, lightweight model-internal mechanism can effectively bridge the gap between corpus retrieval and long-context inference, outperforming specialized SOTA systems on both.

Personal LLM agents can significantly improve recommendation quality by mediating platform rankings with cross-platform history, but naive mediation introduces substantial harmful overrides.

Sequential recommenders can dramatically boost catalog diversity by adaptively pruning action sets, achieving up to 5.21x diversity gains without sacrificing session depth.

A new benchmark reveals that even sophisticated RAG systems struggle with prompt injection, with DistilBERT leading detection at 89.6% F1.

LLM agents can autonomously optimize RAG pipelines by diagnosing retrieval vs. generation failures, outperforming traditional hyperparameter tuning and achieving Pareto efficiency.

A small, causally identified set of attention heads reliably detects network infrastructure information across diverse LLMs, but their responsibility may reside in single neurons or be distributed across the head depending on the model.

Distributed MMD with Random Fourier Features achieves near-perfect correlation with expected drift magnitude on massive e-commerce data, while per-dimension KS tests falter.

Discover hierarchical cluster structure across multiple scales in directed graphs without knowing the number of clusters beforehand, even across sparse bridges.

A novel cross-domain recommendation framework achieves state-of-the-art book-to-music mood matching by fusing sentiment embeddings with LLM-driven emotional quadrant classification.

A single token can bypass most safety filters, revealing a critical vulnerability in current LLM alignment techniques.

LLM inference cost plummets as a novel output-aware KV cache retrieval method bypasses attention score approximation errors, achieving competitive performance with minimal overhead.

Generative auctions can now achieve externality-aware whole-feed optimization with exact strategyproofness using a transformer, overcoming a critical deployment barrier.

Personalized suicide risk forecasting is significantly improved by embedding patients in a latent space that models similarity, outperforming pooled and per-patient models by up to 30% for first-occurrence event prediction.

NeuroCBIR achieves near-perfect subject re-identification in 3D brain MRIs and encodes clinically relevant information for downstream tasks, all within seconds per scan.