A machine-learned router predicts per-query recall for filtered ANN methods and selects the recall-QPS optimal one, outperforming fixed baselines on five unseen datasets.
hub
Milvus: A purpose-built vector data management system
14 Pith papers cite this work, alongside 316 external citations. Polarity classification is still indexing.
14
Pith papers citing it
316
external citations · external index
hub tools
citation-role summary
background 3
dataset 1
citation-polarity summary
representative citing papers
BlobShuffle reduces shuffling costs by over 40x in Kafka Streams by using object storage for batching with notifications, achieving sub-2s 95th-percentile latency and scaling beyond 2 GiB/s.
PERMA is a new benchmark using temporally ordered events, text variability, and linguistic alignment to evaluate LLM memory agents on persona consistency beyond simple retrieval.
Large-scale HPC evaluation of Qdrant, Milvus, and Weaviate reveals that workload patterns limit scaling and extra cores can reduce throughput, exposing a cloud-to-HPC design mismatch.
HKVM-RAG uses key-value-separated hypergraphs to organize LLM evidence tuples into answer-path hyperedges, yielding F1 gains over KG-PPR on two multi-hop QA benchmarks and further gains when combined with dense retrievers.
ANN search quality is better assessed by 1/Ratio@k than Recall@k because the former tracks downstream task utility more closely while allowing substantially lower computational cost.
PipeANN-Filter improves filtered vector search latency and throughput on SSD by exploring a superset of valid vectors identified via probabilistic filters and verifying attributes only after selecting top-k candidates.
FAVOR achieves 1.3-5x higher QPS at 95% Recall@10 for arbitrary filtered ANNS by combining exclusion-distance reshaping in HNSW graphs with a selectivity-driven router that switches between brute-force and optimized search.
TensorHub uses Reference-Oriented Storage to enable scalable weight transfer in LLM RL training by referencing replicated GPU weights, achieving up to 19x reduction in cross-datacenter stall time.
RACORN-1 adds adaptive search fallback to ACORN-1 to fix recall collapse at low selectivity in filtered vector search, achieving 9-26x speedups over HNSW with recovered recall on 1M-40M datasets.
Proposes Governed Evolving Memory (GEM) as a state-trajectory workload for long-term AI agent memory using four operators and six correctness conditions that record-level systems cannot satisfy.
PDR is a user-context-aware framework for LLM research agents that improves report relevance over static baselines, supported by a new dataset and hybrid evaluation.
Relational engines achieve faster SQL+vector-search queries on GPU than CPU when using compact vector indexes and fast interconnects, reversing the CPU-only design in current systems.
This survey discusses key components and challenges for Personal LLM Agents and reviews solutions for their capability, efficiency, and security.
citing papers explorer
-
Query-aware Routing for Filtered Approximate Nearest Neighbors Search
A machine-learned router predicts per-query recall for filtered ANN methods and selects the recall-QPS optimal one, outperforming fixed baselines on five unseen datasets.
-
BlobShuffle: Cost-Effective Repartitioning in Stream Processing Systems via Object Storage Exemplified with Kafka Streams
BlobShuffle reduces shuffling costs by over 40x in Kafka Streams by using object storage for batching with notifications, achieving sub-2s 95th-percentile latency and scaling beyond 2 GiB/s.
-
PERMA: Benchmarking Personalized Memory Agents via Event-Driven Preference and Realistic Task Environments
PERMA is a new benchmark using temporally ordered events, text variability, and linguistic alignment to evaluate LLM memory agents on persona consistency beyond simple retrieval.
-
When More Cores Hurts: The Vector Database Scaling Paradox in HPC
Large-scale HPC evaluation of Qdrant, Milvus, and Weaviate reveals that workload patterns limit scaling and extra cores can reduce throughput, exposing a cloud-to-HPC design mismatch.
-
HKVM-RAG: Key-Value-Separated Hypergraph Evidence Organization for Multi-Hop RAG
HKVM-RAG uses key-value-separated hypergraphs to organize LLM evidence tuples into answer-path hyperedges, yielding F1 gains over KG-PPR on two multi-hop QA benchmarks and further gains when combined with dense retrievers.
-
ANN Search: Recall What Matters
ANN search quality is better assessed by 1/Ratio@k than Recall@k because the former tracks downstream task utility more closely while allowing substantially lower computational cost.
-
PipeANN-Filter: An Efficient Filtered Vector Search System on SSD
PipeANN-Filter improves filtered vector search latency and throughput on SSD by exploring a superset of valid vectors identified via probabilistic filters and verifying attributes only after selecting top-k candidates.
-
FAVOR: Efficient Filter-Agnostic Vector ANNS Based on Selectivity-Aware Exclusion Distances
FAVOR achieves 1.3-5x higher QPS at 95% Recall@10 for arbitrary filtered ANNS by combining exclusion-distance reshaping in HNSW graphs with a selectivity-driven router that switches between brute-force and optimized search.
-
TensorHub: Scalable and Elastic Weight Transfer for LLM RL Training
TensorHub uses Reference-Oriented Storage to enable scalable weight transfer in LLM RL training by referencing replicated GPU weights, achieving up to 19x reduction in cross-datacenter stall time.
-
RACORN-1: Adaptive Recall-Preserving Speedup for Low-Selectivity Filtered Vector Search
RACORN-1 adds adaptive search fallback to ACORN-1 to fix recall collapse at low selectivity in filtered vector search, achieving 9-26x speedups over HNSW with recovered recall on 1M-40M datasets.
-
Is Agent Memory a Database? Rethinking Data Foundations for Long-Term AI Agent Memory
Proposes Governed Evolving Memory (GEM) as a state-trajectory workload for long-term AI agent memory using four operators and six correctness conditions that record-level systems cannot satisfy.
-
Personalized Deep Research: A User-Centric Framework, Dataset, and Hybrid Evaluation for Knowledge Discovery
PDR is a user-context-aware framework for LLM research agents that improves report relevance over static baselines, supported by a new dataset and hybrid evaluation.
-
To GPU or Not to GPU: Vector Search in Relational Engines
Relational engines achieve faster SQL+vector-search queries on GPU than CPU when using compact vector indexes and fast interconnects, reversing the CPU-only design in current systems.
-
Personal LLM Agents: Insights and Survey about the Capability, Efficiency and Security
This survey discusses key components and challenges for Personal LLM Agents and reviews solutions for their capability, efficiency, and security.