StreamKL is the first fused GPU primitive for attention KL divergence that reduces memory from O(N_Q N_K) to O(1) via an online one-pass formulation and tile-wise recomputation.
hub
MINILM: Deep Self-Attention Distillation for Task- Agnostic Compression of Pre-Trained Transformers
28 Pith papers cite this work. Polarity classification is still indexing.
28
Pith papers citing it
hub tools
citation-role summary
background 3
method 1
citation-polarity summary
representative citing papers
A Gaussian information-gain metric in embedding space quantifies semantic progress in dialogues via uncertainty reduction and shows competitive agreement with human judgments on MT-Bench and UltraFeedback.
Face-Feature Tuning is a label-free logit remapping method that reduces FPR/TPR gaps across groups in deepfake detection while preserving overall accuracy.
AuDisAgent reformulates multimodal controversy detection as a dynamic audience dissemination process using screening, panel discussion, and arbitration agents, plus comment bootstrapping, and reports outperforming prior static methods on a public dataset.
ConvMemory v3 introduces a dual-evidence gate for target-conditioned memory validity verification, reporting 90.12% accuracy on synthetic benchmarks, 98.8% transfer to real data, and H@1 improvement from 45.1% to 95.7% in deployment while preserving safety contracts.
An end-to-end SLU architecture with frozen SSL acoustic encoder, LSTM classification head, and cross-modal distillation achieves 93% accuracy on simple commands and 82% on spontaneous speech at 7 ms latency on the new VoiceStick corpus, outperforming cascade baselines.
Graph2Idea builds dynamic knowledge graphs from retrieved literature to supply compact, relational contexts that guide LLMs in generating novel, feasible, and high-quality scientific ideas, outperforming flat-text baselines on automatic metrics.
A knowledge-inclusive PINN framework integrates metagenomics literature and network structure with gLV equations to model microbial interactions, achieving up to 53% improvement over prior methods.
LLM routers across 21 methods on 5 benchmarks converge to similar accuracy below oracle due to learning global performance trends rather than fine-grained query signals.
ImproBR combines a hybrid detector with GPT-4o mini and RAG to raise bug report structural completeness from 7.9% to 96.4% and executable steps from 28.8% to 67.6% on 139 Mojira reports.
Methods for constructing Hypergraphs of Text are proposed with a new effort ratio metric where TF-IDF baselines match LLM methods in experiments.
ORPHEAS, a Greek-English embedding model created with knowledge graph fine-tuning, outperforms state-of-the-art multilingual models on monolingual and cross-lingual retrieval benchmarks.
QTyBERT matches or exceeds BERT-based log anomaly detection effectiveness while reducing embedding generation time to near static word embedding levels.
ERM-based PU classifiers designed for case-control sampling deteriorate under single-sample scenarios, requiring a change in the empirical risk definition; a single-sample analogue of the non-negative risk classifier is introduced and shown to differ notably when many positives are labeled.
Contrastive pre-training on unsupervised data at scale creates text and code embeddings that set new state-of-the-art results on classification and semantic search benchmarks.
PSCT-Net introduces a geometry-aware neural framework that uses differentiable back-projection and attention-guided 3D refinement to reconstruct pediatric skull CT from bi-planar X-rays.
EvoGens uses rank-based mutation, semantic-aware crossover, and lightweight evaluation to evolve populations of LLM-generated scientific ideas, boosting novelty and diversity metrics.
SemStruct models tables as heterogeneous graphs with GNNs on frozen PLM embeddings to incorporate row co-occurrences for schema matching and reports SOTA results on Valentine and SOTAB-SM benchmarks.
Widthwise pruning of LVLM language backbones combined with supervised finetuning and hidden-state distillation recovers over 95% performance using just 5% of data across 3B-7B models.
MIPIC trains Matryoshka representations using self-distilled intra-relational alignment and progressive information chaining, yielding competitive results on STS, NLI, and classification tasks especially at low dimensions.
A 133M-parameter ensemble of fine-tuned mpnet and MiniLM encoders achieves 83.5% accuracy on a 200-task synthetic benchmark for robot skill prediction, beating several larger zero-shot LLMs.
A multi-agent multimodal system with fact-grounded adjudication and a dynamic two-tier preference graph cuts false positives in content filtering by 74.3% and nearly doubles F1-score versus text-only baselines while supporting user-driven Delta adjustments.
A minimalist retrieval-and-generation framework using turn isolation and query-driven pruning outperforms complex memory systems by directly addressing signal sparsity and dual-level redundancy in dialogues.
Sentence transformers show partial zero-shot ability to link route descriptions with hiking queries, indicating some grasp of quasi-geospatial concepts like type and difficulty.
citing papers explorer
-
Text and Code Embeddings by Contrastive Pre-Training
Contrastive pre-training on unsupervised data at scale creates text and code embeddings that set new state-of-the-art results on classification and semantic search benchmarks.