BT-SR uses Barlow Twins decorrelation as a control knob trading head-item accuracy for long-tail exposure
_reachsumit · x · 2026-09-22
A paper proposes BT-SR, a decorrelation-regularized framework for controllable sequential recommendation.
- Idea: augment next-item prediction with a Barlow Twins-style redundancy-reduction term; positive pairs are user histories sharing the same next-item target, avoiding synthetic corruptions.
- Geometric analysis: decorrelation suppresses shared low-rank directions that give popular items a global scoring advantage, with a bucket-based alignment-concentration metric to quantify the effect.
- Results: consistent next-item ranking gains across five benchmarks; decorrelation strength acts as a control knob reallocating accuracy between head and tail items, enabling accuracy-exposure trade-offs.
More from Research
- Tencent ARC Lab releases GameHorizon Suite to benchmark VLMs and agents on AAA gameplay — yshan2u · 2026-09-22
- Stanford turns papers into AI agents; two unrelated studies surface unreported ADHD genetic link — VraserX · 2026-09-22
- Xiaomi MiMo 2.6 Pro formalizes Li–Yorke chaos theorem in 6,000+ lines of verified Lean — bookwormengr · 2026-09-22
- Apollo, the first advanced LLM for Ancient Greek, aims to restore tattered papyri — nordicinst · 2026-09-22
- Two Minute Papers: Jev claims 200x AI speedup via System One models — but there's a catch — Two Minute Papers · 2026-09-22
- New LLM fills gaps in tattered ancient Greek papyrus records — Wired AI · 2026-09-22