fusion

This paper proposes MIDAS, a unified framework for incomplete multimodal sentiment analysis that uses mutual information disentanglement and uncertainty-aware fusion to robustly represent and integrate modalities under missing-data conditions.

A new sub-6B sparse activation AI model built with a fusion architecture combines weights from LFM2.5-2.6B and Qwen3.6-35B-A3B, achieving near-Qwen3.6-35B performance at a fraction of the size.

China has constructed a 582-ton giant magnet to support its 'Artificial Sun' fusion reactor, a significant step in nuclear fusion research.

TechCrunch Disrupt 2026 will feature a Smart Systems Stage focusing on AI infrastructure challenges, including fusion energy, grid modernization, and data center power demands. Leaders from Commonwealth Fusion Systems, Helion, Inertia, Bloom Energy, and others will speak.

A research paper proposing SeRIn, a multimodal fusion scheme that separates modality-specific refinement from cross-modal interaction, achieving state-of-the-art on CH-SIMS and CMU-MOSEI benchmarks for sentiment analysis.

OmniPM-Net is a neural process fusion model that combines discrete station forecasts from graph neural networks with gridded forecasts from chemical transport models to produce consistent PM10 predictions at both stations and grid cells, improving accuracy especially during dust storms.

Meta introduces techniques like Lazy Pre-Norm, Multi-CTA Norm Fusion, and FlashNormAttention to fuse normalization operations with GEMM and Attention kernels, hiding up to 90% of normalization latency on NVIDIA B200 hardware and achieving up to 35% latency reduction in attention blocks.

This paper proposes SHAP-weighted cross-modal expert fusion (XGAF) for emotion and sentiment recognition, demonstrating that sum-abs SHAP aggregation achieves early-fusion-level performance on MELD and CMU-MOSEI datasets.

This paper investigates the use of conversational temporal dynamics (turn-pair timing) as a lightweight modality for automatic depression detection from dyadic clinical interviews, showing that a compact 24-dimensional timing module achieves strong performance and complements standard acoustic and semantic features when fused.

A tweet celebrating American technological achievements from the transistor to the internet, asserting that fusion energy is the next leap and is being built in the US, on Independence Day.

The author built an open-source agent that uses a panel of different LLMs with a judge and synthesizer for hard reasoning steps, alongside cost-aware routing, layered memory, governance, and subagent support. It is alpha software with mixed benchmarks on fusion effectiveness.

Adam CAD Copilot is an AI assistant for CAD design, integrated into Onshape and Fusion.

OpenRouter's Fusion API offers pricing and provider information for routing AI model requests across multiple providers, enabling flexible and cost-effective access to various AI models.

This paper applies the likelihood ratio framework for forensic authorship attribution to Japanese texts, fusing stylometric features with embedding-based systems to improve discrimination and calibration.

Open Router announced Fusion, a new system that may be similar to OpenAI's pro model, generating excitement in the community.

OpenRouter introduces the Fusion API, a compound AI model that achieves Fable-level intelligence at half the price.

This blog post continues the profiling in PyTorch series, exploring nn.Linear, MLP blocks, and fusion techniques using Triton kernels to optimize performance.

This article explains how continuation-passing style (CPS) can be used to fuse database operators, avoiding materialization of intermediate results and improving performance, as an alternative to vectorization and compilation. The author presents a simplified example with numeric list operations and describes the inspiration from a short paper.

This paper introduces a plug-in calibration module that adjusts multimodal representations before fusion, using cross-modal context to suppress misleading signals and emphasize reliable ones, improving performance on multiple benchmarks.

This paper presents a formal roadmap for transitioning from late-fusion multimodal approaches to native multimodal modeling (NMM) within a unified transformer framework, categorizing existing models by input-output duality and systematically addressing architectural coordination, data curation, training recipes, and evaluation.