「retrieval」の検索結果
53 件Answering geographic queries over Earth observation (EO) archives requires solving a joint retrieval-and-deriv
Retrieval-augmented Web applications increasingly employ multi-agent debate (MAD) to aggregate conflicting, no
External text can affect a retrieval-augmented generation (RAG) system only after the retriever admits it to t
Power-grid customer service increasingly relies on community social media, where customers describe outages, b
Mainstream sequence-processing AI is temporally reactive: each update is driven by input and learned state tra
Why do language models fail on long inputs? The common answer is length.
Large foundation models often answer even when the available visual or textual evidence is insufficient to sup
LLM agents for GPU kernel optimization sit at two ends of a structure-freedom tradeoff. Scripted agents fix wh
On July 1, 2025, Vietnam reduced its 63 provinces to 34, abolished its 696 districts, and replaced its 10,035
We classify real and complex state-space models (SSMs) with linear transitions, distinguishing fixed-basis dia
Claim verification is the task of deciding whether a source text supports a claim, contradicts it, or lacks su
Generative search engines increasingly shape how users discover products, services, and commercial entities by
A stored experience may answer one future question and offer little help with another. We measure this task-co
Next-visit diagnosis prediction forecasts the diagnoses documented at a patient's next encounter, turning retr
Connecting new generation and demand to the grid depends on two scarce skills: checking applications against a
A memory-augmented agent must decide which memories to keep, merge and retrieve, and its only evidence of thei
Assessing regulatory compliance for high-risk AI systems requires systematic identification of technical gaps
Remote sensing image fusion and semantic interpretation are usually optimized separately, so reconstruction qu
Reclamation of saline-alkali land requires management plans that account for soil and water conditions, treatm
Law changes over time. A court decides a case under the law in force on that day, and legal QA benchmarks grad
Previous research shows that vision-language (VL) models struggle to understand fine-grained image details and
Dhārā Bench is a retrieval benchmark for Indian Knowledge Systems (IKS), with 325 questions in 12 categories a
Scaling depth is a key route to stronger large language models, but Pre-LN architectures often suffer from dep
Complex documents such as scientific papers contain rich information in tables, figures, and equations; howeve
This paper studies retrieval-augmented generation (RAG), which leverages external knowledge to improve the per
We present ICJ-Graph, a machine-readable resource connecting documents of the International Court of Justice (
Enterprise conversational systems require memory to personalize interactions, retain information across sessio
Large language models are increasingly consulted for religious knowledge, including the recall of specific Had
Attention-based rerankers score each candidate by the attention mass that retrieval heads route to it from the
Long-term conversational memory faces a timing mismatch: write-time summaries and graphs organize history befo
Retrieval-augmented generation (RAG) extends large language models with external evidence, yet relevant eviden
Multimodal retrieval must integrate complementary evidence from views that may be missing, noisy, or unequally
Cloth-changing person re-identification (CC-ReID) is still very challenging, and no method clearly dominates:
3-D road anomaly segmentation is essential for risk-aware perception in autonomous driving. Its central challe
Clinical evidence is inherently relational: medical images are interpreted with prior studies, cross-modal exa
In virtual environments, choosing appropriate outfits requires contextual harmony with surrounding scenes, yet
Pre-touch sensing, or sensing information about an object before contact, is an attractive modality for robot
Robust multimodal perception for soft robotic grasping is challenged by modality-dependent noise, semantic amb
Key-value (KV) cache eviction methods rank tokens by attention scores, and each layer scores its own cache. To
Social-media comments are interactions, yet they are often translated and evaluated as isolated text. This mis
Social media translation requires preserving meaning while capturing the cultural nuances of user-generated co
Self-evolving agents accumulate experience from task streams, but existing frameworks typically rely on fixed
Anticipating financial outcomes requires synthesizing evolving evidence and reasoning under uncertainty, makin
Humans continuously acquire knowledge from carefully crafted materials such as books and tutorial videos, carr
Purpose: Retrieval-augmented generation (RAG) can introduce electrocardiogram (ECG) criteria that lack patient
Retrieving objects from densely packed storage requires anticipating how contact disturbs surrounding objects
Long-horizon embodied manipulation requires policies to act on information that is no longer visible. Modified
Retrieval-augmented generation (RAG) systems have seen growing adoption in embodied settings as language-query
Modern data analysis pipelines increasingly rely on large datasets assembled through retrieval, scraping, auto
Formula-Driven Supervised Learning (FDSL) enables vision pre-training without real images, but existing method
Spiking neural networks (SNNs) have increasingly incorporated rich neuronal dynamics to enhance temporal proce
Dasgupta, Stevens and Navlakha (2017) showed that the Drosophila olfactory circuit, modelled as a random spars
Every day, people continuously infer situational context and adjust the way they understand and remember the w