Web This site Trusted
16 result(s) for “okapi bm25 site hotchips pp site dlmultimedia epa in site can kernel org site:arxiv.org” · 647 ms
Scope: Nishi’s own index (our crawl + Common Crawl) — not the whole live web. Some sites block all automated crawlers (anti-bot walls), so their pages may be absent here even though we never filter lawful content.
Freshness: newest result fetched 24 d ago · median age of these results 24 d · 1 of 16 carry a fetch time
The query asks the model to recall the topic of an earlier conversation about Tesla; the gold answer is anchored in one specific session about Pooja’s dig site
from the top result below
https://arxiv.org/html/2608.02613v1
A benchmark for personal memory in this setting should therefore evaluate the full stack on device, not just whether a model can retrieve the relevant information
https://arxiv.org/html/2501.13956v1
com
Abstract
We introduce Zep, a novel memory layer service for AI agents that outperforms the current state-of-the-art system, MemGPT, in the Deep Memory Retrieval (DMR) benchmark
https://arxiv.org/html/2608.03048v1
for large language models, as recent recurrent-memory approaches face two inherent challenges: sequential chunk-wise updates can overwrite early critical evidence with later irrelevant content, and serial inter-chunk dependencies limit parallelism and cause latency to increase with context length
https://arxiv.org/list/cs.AI/new
In practice, many field descriptions remain ambiguous or incomplete, as much of the essential context (e
https://arxiv.org/pdf/2608.03411
{~l F )=PH# ӑHton> stream DUD: Decoupled Update Dynamics for Reliable Uncertainty Quantification in Large Language ModelsYixin BuRunze XiaGuanyun ZouYupeng JiHaodong LiuPiji Lihttp://creativecommons
https://arxiv.org/list/cs.CL/new
LG)
Half a million cuneiform clay tablets survive in museums worldwide, yet modern users can neither read nor write in the world's oldest writing system, leaving a 4,000-year cultural barrier that existing NLP tools have only partially addressed
https://arxiv.org/abs/2608.03860
In contrast, an MS MARCO-trained cross-encoder reranker reduces precision on the scientific corpus, suggesting that domain mismatch can outweigh the benefits of stronger query-passage interaction
https://arxiv.org/abs/2608.03883
All submitted systems follow a staged retrieval design in which BM25 retrieves candidates at scale and semantic or LLM rerankers refine the final rankings
https://arxiv.org/list/cs.IR/new
Our results show that skill representation itself matters, and that simply preserving the structure already present in skill files can substantially improve retrieval
https://arxiv.org/list/cs.IR/recent
IR)
Tue, 4 Aug 2026 (showing first 34 of 45 entries )
[17] arXiv:2608
https://arxiv.org/abs/2104.08663
Benchmark for Zero-shot Evaluation of Information Retrieval Models, by Nandan Thakur and 4 other authors
View PDF HTML (experimental)
Abstract:Existing neural information retrieval (IR) models have often been studied in homogeneous and narrow settings, which has considerably limited insights into their out-of
https://arxiv.org/html/2407.16741v3
In particular, AI agents, systems that can perceive and act upon the external environment, have recently received ever-increasing research focus
https://arxiv.org/pdf/2608.03358
tbwpә؎/Iܸ:Ai[,&(%;nf,(^Xp)"[" (EPA(ܺR(%Jhʠ j%ß B-PeRheneᴨ&ve(+2ƍ@kP3&KQO"-G oՔ22
https://arxiv.org/pdf/2510.03215
7 % 1 0 obj > endobj 2 0 obj > endobj 3 0 obj > stream Cache-to-Cache: Direct Semantic Communication Between Large Language ModelsTianyu FuZihan MinHanling ZhangJichao YanGuohao DaiWanli OuyangYu Wanghttp://arxiv
https://arxiv.org/pdf/2607.26327v3
'-~} HZ$ {-0B/pF '|HBI-EJ+-0$sJ( #,5BKZb' 0v A ,:i82 #Mh0i-aMB[aVF 0 K 3J8F8 t 9 2a5Qa}.!PIJp^J 6;Z# Nea>6 EpYJ
https://arxiv.org/pdf/2407.00215
stream xP( endstream endobj 11 0 obj > stream xP( endstream endobj 17 0 obj > stream xP( endstream endobj 20 0 obj > stream xP( endstream endobj 23 0 obj > stream xP( endstream endobj