←Search results

Search “Multimodal”

By relevance · Total 801 results Clear search

Wednesday

  1. FactRelated story

    Miles summarizes the eight-step multimodal data cleaning solution for FDE at beauty companies

    On October 28, 2026, Miles summarized an eight-step multimodal data cleaning solution for FDE implementation projects at beauty companies. The solution processes text, image, audio, and video data collected from Xiaohongshu, Douyin, WeChat Channels, and Weibo through eight steps: defining data specifications, preserving original content, mapping data relationships, multimodal normalization, conflict and quality cleaning, generating structured data, performing acceptance checks, and storing data by rules and versions. The original text emphasizes that cleaning results must not overwrite the original text, and that OCR, ASR, and model labels serve only as newly added results; when ASR and OCR values conflict, the original evidence is traced back, and if it cannot be confirmed, the item enters a pending-verification queue. This pipeline has also been organized into a general solution that can be reused in scenarios such as consumer goods insights, automotive word-of-mouth, restaurant site selection, and brand public opinion.

    Matched fieldsTitle

    View story:Miles summarizes the eight-step multimodal data cleaning solution for FDE at beauty companies

Friday

  1. Story1 reports · 1 facts

    Zhang Changqing presents a report on generalized multimodal fusion technology at CNCC 2026

    On October 21, 2026, Zhang Changqing of Tianjin University gave a talk titled "Generalized Multimodal Fusion Technology" at CNCC 2026, presenting the team's research on multimodal representation, dynamic trustworthy fusion, and the general-purpose prediction foundation model Fusion Anything Model (FAM). According to an October 28 report by Leiphone, its TMC method uses evidence theory to estimate uncertainty, raising crystal structure recognition accuracy from about 58%–59% to around 93.9%; FAM was tested on 18 real-world datasets, 12 modalities, and 11 types of prediction tasks, with average accuracy improving by about 6.9 percentage points over specialized fusion baselines. Currently, the content of this talk has been disclosed by a single source, with no subsequent developments.

    On October 21, 2026, Zhang Changqing of Tianjin University gave a talk titled "Generalized Multimodal Fusion Technology" at CNCC 2026, presenting the team's research on multimodal representation, dynamic trustworthy fusion, and the general-purpose prediction foundation model Fusion Anything Model (FAM). According to an October 28 report by Leiphone, its TMC method uses evidence theory to estimate uncertainty, raising crystal structure recognition accuracy from about 58%–59% to around 93.9%; FAM

    Matched fieldsTitle

Wednesday

  1. FactRelated story

    Zhang Changqing presents a report on generalized multimodal fusion technology at CNCC 2026

    On October 21, 2026, Zhang Changqing of Tianjin University gave a talk titled "Generalized Multimodal Fusion Technology" at CNCC 2026, presenting the team's research on multimodal representation, dynamic trustworthy fusion, and the general-purpose prediction foundation model Fusion Anything Model (FAM). According to an October 28 report by Leiphone, its TMC method uses evidence theory to estimate uncertainty, raising crystal structure recognition accuracy from about 58%–59% to around 93.9%; FAM was tested on 18 real-world datasets, 12 modalities, and 11 types of prediction tasks, with average accuracy improving by about 6.9 percentage points over specialized fusion baselines. Currently, the content of this talk has been disclosed by a single source, with no subsequent developments.

    Matched fieldsTitle

    View story:Zhang Changqing presents a report on generalized multimodal fusion technology at CNCC 2026

  2. FactRelated story

    Flatkey gateway supports calling multimodal models in Codex and Claude Code via CLI

    As a unified API gateway, Flatkey supports calling multimodal models in Codex and Claude Code via CLI. According to a report dated October 21, 2026, the gateway can call text models such as GPT, Claude, DeepSeek, GLM, and Kimi, as well as image and video models such as GPT Image and Seedance; users can access 100+ official models and 1,000+ tool APIs with a single Key, with costs consolidated into one bill. The report also states that its everyday prices are at least 20% lower than official pricing; during the October 19–23 promotion, GPT, GLM 5.2, Kimi K3, and DeepSeek are 40% off, Claude is 20% off, and Seedance 2.0 is 15% off. There are currently no subsequent developments.

    Matched fieldsTitle

    View story:The Flatkey gateway supports calling multimodal models in Codex and Claude Code via CLI

Sunday

  1. Story1 reports · 1 facts

    The Flatkey gateway supports calling multimodal models in Codex and Claude Code via CLI

    As a unified API gateway, Flatkey supports calling multimodal models in Codex and Claude Code via CLI. According to a report dated October 21, 2026, the gateway can call text models such as GPT, Claude, DeepSeek, GLM, and Kimi, as well as image and video models such as GPT Image and Seedance; users can access 100+ official models and 1,000+ tool APIs with a single Key, with costs consolidated into one bill. The report also states that its everyday prices are at least 20% lower than official pricing; during the October 19–23 promotion, GPT, GLM 5.2, Kimi K3, and DeepSeek are 40% off, Claude is 20% off, and Seedance 2.0 is 15% off. There are currently no subsequent developments.

    As a unified API gateway, Flatkey supports calling multimodal models in Codex and Claude Code via CLI. According to a report dated October 21, 2026, the gateway can call text models such as GPT, Claude, DeepSeek, GLM, and Kimi, as well as image and video models such as GPT Image and Seedance; users can access 100+ official models and 1,000+ tool APIs with a single Key, with costs consolidated into one bill. The report also states that its everyday prices are at least 20% lower than official pric

    Matched fieldsTitle

  2. Story1 reports · 1 facts

    Luma proposes an exact sequence parallelism method for packed multimodal masking

    Luma proposes an exact sequence parallelism method for packed multimodal masking. At each layer, the method chooses among ring attention, sparse all-to-all, or Ulysses, and its numerical results are consistent with those on a single GPU. On up to 128 H100 GPUs, the method outperforms the FlashMask + Ulysses hybrid scheme in 47 of 48 cases. There are currently no follow-up developments.

    Matched fieldsTitle

  3. FactRelated story

    Luma proposes an exact sequence parallelism method for packed multimodal masks

    Luma proposes an exact sequence parallelism method for packed multimodal masking. At each layer, the method chooses among ring attention, sparse all-to-all, or Ulysses, and its numerical results are consistent with those on a single GPU. On up to 128 H100 GPUs, the method outperforms the FlashMask + Ulysses hybrid scheme in 47 of 48 cases. There are currently no follow-up developments.

    Matched fieldsTitle

    View story:Luma proposes an exact sequence parallelism method for packed multimodal masking

  4. Story1 reports · 1 facts

    Central China Normal University and others propose multimodal medical image fusion method CoReFuse-Med

    Central China Normal University, Huazhong University of Science and Technology, Guangzhou University of Chinese Medicine and its First Affiliated Hospital, and the University of North Carolina at Chapel Hill propose a multimodal MRI segmentation network, CoReFuse-Med. The method uses LACE+IMSH to remove resampling artifacts at shallow layers and MBCF to calibrate modality contributions at deep layers, addressing the problem of low-quality modalities contaminating the fusion. On the EPVS dataset, it achieves a DSC of 0.7326 and an HD95 of 7.35, outperforming MedNeXt's 0.7085 and 8.17; on BraTS, its DSC is 0.8528 with only 2.665M parameters. The method targets the resolution degradation after spatial alignment and does not cover scenarios with missing modalities or strong noise.

    Central China Normal University, Huazhong University of Science and Technology, Guangzhou University of Chinese Medicine and its First Affiliated Hospital, and the University of North Carolina at Chapel Hill propose a multimodal MRI segmentation network, CoReFuse-Med. The method uses LACE+IMSH to remove resampling artifacts at shallow layers and MBCF to calibrate modality contributions at deep layers, addressing the problem of low-quality modalities contaminating the fusion. On the EPVS dataset,

    Matched fieldsTitle

Tuesday

Monday

  3. FactRelated story

    Central China Normal University and others propose multimodal medical image fusion method CoReFuse-Med

    Central China Normal University, Huazhong University of Science and Technology, Guangzhou University of Chinese Medicine and its First Affiliated Hospital, and the University of North Carolina at Chapel Hill propose a multimodal MRI segmentation network, CoReFuse-Med. The method uses LACE+IMSH to remove resampling artifacts at shallow layers and MBCF to calibrate modality contributions at deep layers, addressing the problem of low-quality modalities contaminating the fusion. On the EPVS dataset, it achieves a DSC of 0.7326 and an HD95 of 7.35, outperforming MedNeXt's 0.7085 and 8.17; on BraTS, its DSC is 0.8528 with only 2.665M parameters. The method targets the resolution degradation after spatial alignment and does not cover scenarios with missing modalities or strong noise.

    Matched fieldsTitle

    View story:Central China Normal University and others propose multimodal medical image fusion method CoReFuse-Med

Sunday

Friday

Thursday