Twitxy Saved

Sumit

@_reachsumit

Senior ML Engineer @Meta | prev: @TikTok_us, @Amazon, @Samsung | UChicago Alum

blog.reachsumit.com

๐Ÿ‡ฎ๐Ÿ‡ณโ†’๐Ÿ‡ฐ๐Ÿ‡ทโ†’๐Ÿ‡ฆ๐Ÿ‡บโ†’๐Ÿ‡จ๐Ÿ‡ฆโ†’๐Ÿ‡บ๐Ÿ‡ฒ
Seattle, WA recsys.substack.com Joined April 2010
519 Following 4.5K Followers 11K Posts

Latest posts 5

Pinned post

Sumit @_reachsumit
When Direction Matters: Diagnosing Relational Instruction Following in Dense Retrieval

Shows that dense retrievers follow keyword-based instructions but often fail when an instruction only reverses a relation over the same entities.

๐Ÿ“ arxiv.org/abs/2611.15417
arxiv.org When Direction Matters: Diagnosing Relational Instruction... Dense retrievers increasingly claim the ability to condition retrieval on natural-language relevance criteria beyond a short query. We identify a specific weakness in this capability: the...
Sumit @_reachsumit
REWEAVE: Online Learning for Workload-Adaptive HNSW Indexes

Introduces an online method that learns which HNSW edges matter for the query workload and rewires them within the same per-node degree cap, improving recall on skewed or shifting queries.

๐Ÿ“ arxiv.org/abs/2611.16060
arxiv.org REWEAVE: Online Learning for Workload-Adaptive HNSW Indexes Vector databases rely on Hierarchical Navigable Small World (HNSW) indexes for similarity search, but a graph built from data alone can be poorly matched to a skewed or changing query workload. We...
Sumit @_reachsumit
From Tokens to Behaviors: Rethinking Generative Recommendation for User Trajectory Prediction

Proposes a generative recommender that predicts a user's future behavior sequence, not just the next item.

๐Ÿ“ arxiv.org/abs/2611.16159
๐Ÿ‘จ๐Ÿฝโ€๐Ÿ’ป github.com/CalmaQi/GR4UBT
arxiv.org From Tokens to Behaviors: Rethinking Generative Recommendation for... User behavior trajectory prediction (UBTP) aims to capture how a user's needs evolve over time by predicting the sequence of future interactions from that user's interaction history. Recent...