AI & ML
RecMem: Recurrent Memory Compression for Long-Sequence Generative Recommendation
arXiv cs.IR0 views
arXiv:2611.17921v1 Announce Type: cross Abstract: Generative recommendation benefits from long interaction histories, but the cost of processing them directly grows with history length. We propose RecMem, a recurrent memory compression framework guided by three desiderata: space invariance, semantic fidelity, and lifelong evolvability. RecMem compresses historical segments into a fixed set of memory slots using a bifurcated encoder-decoder with shared lower layers and slot-wise forget-gated updates. Recommendation supervision trains the memory to preserve predictive information, while stream-consistency training aligns predictions from sequential and packed processing. By construction, the memory state and decoder context are independent of history length; under contraction and approximate semigroup consistency, we further derive a horizon-independent bound on the state discrepancy between these two processing orders. On MerRec, RecMem uses 16 memory slots and a 216-token decoder context, compared with 1,200 tokens for HSTU-1200, while retaining 87.9% of HSTU-1200's Recall@10 and 97.0% of its Recall@200. It outperforms PHSTU and the evaluated recent-only and retrieval baselines, and analyses of distant history show that archived interactions contribute predictive signal.
Source:https://arxiv.org/abs/2611.17921PreviousMeReader: An Offline, Privacy-Preserving, Progress-Aware AI Assistant for Narrative eBook ReadingNextStreamEmbed: Real-Time Indexing from Pixels for Screenshot Retrieval
Related posts

Add ChatGPT-style AI chat to your Expo app in 5 minutes
0dev.to OpenSource
sheet_route: نوافذ bottom sheet في Flutter كـ routes حقيقية مع snap points و go_router
0dev.to OpenSource
Every family WhatsApp group in India has the same problem
0dev.to OpenSource

How to Use Git Stash Step by Step
0dev.to Git
GenTUS: Table Union Search via Constrained Generation of Semantic Table Identifiers
0arXiv cs.DB
Larch: Learned Query Optimization for Semantic Predicates
0arXiv cs.DB