Tagged articles
cross-attention
11 articles · Page 1 of 1
Shanghai Institute for Science Intelligence open-sources REMORY, a 1.9B parameter memory module that uses cross-attention to generate residual tokens from compressed context, enabling small models to autonomously conduct multi-day research with 50% fewer tool errors and 33% lower token costs, matching GPT-6.1-Terra performance.
BrowseCompREMORYTerminal-Bench
0 likes · 7 min read
Researchers from Northwestern Polytechnical University and HKUST propose Dynamic Gated Pathways (DGP), a method that lets each LLM layer dynamically attend to all ViT layers, improving multimodal benchmarks by learning per-layer, per-sample routing weights instead of using a fixed visual projection.
DGPDynamic Gated PathwaysLLM
0 likes · 18 min read
This article details QueryFormer, the winning solution for KDD Cup 2026 Tencent UniRec Challenge, which introduces a cross-attention query generation mechanism to unify feature interaction and behavior sequence modeling, achieving 0.83254 AUC on the official test set.
AUCKDD Cup 2026QueryFormer
0 likes · 24 min read
This article introduces three complementary attention heatmap visualizations — image-to-text, text-to-image, and image-to-image — for interpreting text-to-image diffusion models, detailing aggregated attention computation, bilinear interpolation rendering, and an interactive Flask web service to diagnose prompt adherence and model biases.
Flask web serviceSelf-Attentionattention visualization
0 likes · 12 min read
This article explains how to compute and visualize attention heatmaps for text‑to‑image diffusion models, offering three complementary views (image‑to‑text, text‑to‑image, image‑to‑image), detailing the aggregation formulas, rendering process, and an interactive Flask web service that helps diagnose prompt failures and reveal model biases.
Flaskattention visualizationcross-attention
0 likes · 10 min read
This paper introduces a novel framework that constructs and aligns dual visual‑textual trees on heterogeneous hyperbolic manifolds, addressing asymmetric modality alignment in hierarchical classification tasks and achieving state‑of‑the‑art performance on benchmarks such as CIFAR‑100, ImageNet and Rare Species datasets.
cross-attentionhierarchical classificationhyperbolic manifolds
0 likes · 8 min read
The VeM model introduces a latent diffusion framework that leverages hierarchical video parsing, scene‑guided cross‑attention, and a transition‑beat alignment adapter to generate high‑fidelity background music perfectly synchronized with video semantics, timing, and rhythm, outperforming existing baselines on extensive quantitative and qualitative evaluations.
Latent DiffusionTemporal Alignmentaudio generation
0 likes · 14 min read
This article surveys the latest multimodal large language model research, dissecting the design, training strategies, and performance trade‑offs of models such as Llama 3.2, Molmo, NVLM, Qwen2‑VL, Pixtral, MM1.5, Emu3, and Janus, and highlights the challenges of fair cross‑model evaluation.
AI researchLarge Language ModelsModel Training Strategies
0 likes · 16 min read
This article explains the fundamentals of multimodal large language models, covering their definition, typical applications, two main architectural approaches—unified embedding decoder and cross‑modal attention—along with detailed component breakdowns, a PyTorch implementation of image‑patch projection, and training considerations, ending with a discussion of trade‑offs between the methods.
Image EncoderLinear ProjectionPyTorch
0 likes · 14 min read
NewBeeNLPMay 9, 2024 · Artificial Intelligence
This article summarizes the UniSAR framework, which models four types of fine‑grained user transitions between search and recommendation, demonstrates its effectiveness on public datasets, and shows how joint learning improves both search relevance and recommendation quality.
TransformerUniSARcross-attention
0 likes · 4 min read
The article demystifies Stable Diffusion by explaining its low‑cost latent‑space design and conditioning mechanisms, comparing it to autoregressive, VAE, flow‑based and GAN models, detailing the iterative noise‑to‑image process, token‑based text‑to‑image control, version differences, common generation issues, and providing implementation code examples.
AI image generationComputer VisionStable Diffusion
0 likes · 15 min read
