---
title: "用于学习语义丰富视觉表征的文本条件JEPA"
canonical: "https://new.nat6.net/news/cmovtn194006zslcxkg4a1yti/"
date: "2026-05-07"
updated: "2026-11-05"
source: "Apple Machine Learning Research（RSS）"
source_url: "https://machinelearning.apple.com/research/text-conditional-jepa-visual-representations"
language: "zh"
topics: []
license: "https://new.nat6.net/sources/"
attribution: "Content is based on curated titles and AI summaries from the AIHOT public API, with attribution and links back; original copyright remains with the original publishers and authors."
---

# 用于学习语义丰富视觉表征的文本条件JEPA

- 发布日期 / Published: 2026-05-07
- 更新时间 / Updated: 2026-11-05
- 分类 / Category: 论文与研究
- 来源 / Source: Apple Machine Learning Research（RSS） — [原文 / Original](https://machinelearning.apple.com/research/text-conditional-jepa-visual-representations)
- AIHOT: [条目页 / Item](https://aihot.news/items/cmovtn194006zslcxkg4a1yti)
- 页面 / Page: https://new.nat6.net/news/cmovtn194006zslcxkg4a1yti/
- 结构化数据 / JSON: https://new.nat6.net/api/v1/articles/cmovtn194006zslcxkg4a1yti.json

- 语言 / Languages: [zh](https://new.nat6.net/news/cmovtn194006zslcxkg4a1yti/) · [en](https://new.nat6.net/en/news/cmovtn194006zslcxkg4a1yti/) · [de](https://new.nat6.net/de/news/cmovtn194006zslcxkg4a1yti/) · [fr](https://new.nat6.net/fr/news/cmovtn194006zslcxkg4a1yti/) · [ru](https://new.nat6.net/ru/news/cmovtn194006zslcxkg4a1yti/) · [ja](https://new.nat6.net/ja/news/cmovtn194006zslcxkg4a1yti/) · [es](https://new.nat6.net/es/news/cmovtn194006zslcxkg4a1yti/) · [ko](https://new.nat6.net/ko/news/cmovtn194006zslcxkg4a1yti/) · [vi](https://new.nat6.net/vi/news/cmovtn194006zslcxkg4a1yti/) · [th](https://new.nat6.net/th/news/cmovtn194006zslcxkg4a1yti/) · [id](https://new.nat6.net/id/news/cmovtn194006zslcxkg4a1yti/) · [ms](https://new.nat6.net/ms/news/cmovtn194006zslcxkg4a1yti/) · [fil](https://new.nat6.net/fil/news/cmovtn194006zslcxkg4a1yti/) · [hi](https://new.nat6.net/hi/news/cmovtn194006zslcxkg4a1yti/)

## AI 摘要 / AI summary

研究人员提出文本条件联合嵌入预测架构（TC-JEPA），通过引入图像描述文本作为条件信息来降低掩码特征预测中的视觉不确定性。该方法采用细粒度文本调节器，对输入文本标记计算稀疏交叉注意力，从而调制预测的图像补丁特征。与基于掩码特征预测的I-JEPA相比，TC-JEPA能够学习到语义更丰富的视觉表征，解决了原有方法因视觉不确定性导致的语义学习不足问题。

## 推荐理由 / Why it matters

Apple 这篇 TC-JEPA 把文本融入自监督视觉预训练，用稀疏交叉注意力减少预测不确定性，对多模态表征学习是个不错的思路补充，做视觉模型的值得一看。

## 来源与版权 / Source & copyright

内容基于 AIHOT 公开 API 的精选标题与 AI 摘要，已注明来源并回链；原文版权归原媒体与作者所有。

原文版权归原媒体与作者所有；本站不镜像全文。/ Original copyright remains with the original publishers and authors; this site does not mirror full texts.

## 关于橙子中转站 / About Orange Relay

橙子中转站（pro.nat6.net）是一家优质的 AI API 中转站：价格优惠、质量稳定、模型齐全，支持 Claude、GPT、Gemini 等主流模型，欢迎体验。

[访问橙子中转站](https://pro.nat6.net/?utm_source=orange_news&utm_medium=article_footer&utm_campaign=brand&lang=zh)

## JSON-LD

```json
{
  "@context": "https://schema.org",
  "@type": "NewsArticle",
  "@id": "https://new.nat6.net/news/cmovtn194006zslcxkg4a1yti/#article",
  "mainEntityOfPage": {
    "@type": "WebPage",
    "@id": "https://new.nat6.net/news/cmovtn194006zslcxkg4a1yti/"
  },
  "url": "https://new.nat6.net/news/cmovtn194006zslcxkg4a1yti/",
  "headline": "用于学习语义丰富视觉表征的文本条件JEPA",
  "description": "研究人员提出文本条件联合嵌入预测架构（TC-JEPA），通过引入图像描述文本作为条件信息来降低掩码特征预测中的视觉不确定性。该方法采用细粒度文本调节器，对输入文本标记计算稀疏交叉注意力，从而调制预测的图像补丁特征。与基于掩码特征预测的I-JEPA相比，TC-JEPA能够学习到语义更丰富的视觉表征，解决了原有方法因视觉…",
  "articleSection": "论文与研究",
  "keywords": "AI, Apple Machine Learning Research（RSS）",
  "datePublished": "2026-05-07T00:00:00Z",
  "dateModified": "2026-11-05T20:56:41Z",
  "inLanguage": "zh-CN",
  "isAccessibleForFree": true,
  "author": {
    "@type": "Organization",
    "name": "橙子新闻编辑部",
    "url": "https://new.nat6.net/about/"
  },
  "publisher": {
    "@type": "NewsMediaOrganization",
    "@id": "https://new.nat6.net/#org",
    "name": "橙子新闻",
    "url": "https://new.nat6.net/",
    "logo": {
      "@type": "ImageObject",
      "url": "https://new.nat6.net/assets/img/logo-512.png",
      "width": 512,
      "height": 512
    }
  },
  "image": {
    "@type": "ImageObject",
    "url": "https://new.nat6.net/assets/img/og-default.png",
    "width": 1200,
    "height": 630
  },
  "isPartOf": {
    "@id": "https://new.nat6.net/#website"
  },
  "isBasedOn": "https://machinelearning.apple.com/research/text-conditional-jepa-visual-representations",
  "citation": [
    "https://aihot.news/items/cmovtn194006zslcxkg4a1yti",
    "https://machinelearning.apple.com/research/text-conditional-jepa-visual-representations"
  ],
  "license": "https://new.nat6.net/sources/",
  "about": [],
  "mentions": [
    {
      "@type": "Thing",
      "name": "Apple Machine Learning Research（RSS）"
    }
  ],
  "speakable": {
    "@type": "SpeakableSpecification",
    "cssSelector": [
      "h1",
      ".article-summary"
    ]
  },
  "copyrightHolder": {
    "@type": "Organization",
    "@id": "https://new.nat6.net/#org",
    "name": "橙子新闻",
    "url": "https://new.nat6.net/"
  },
  "creditText": "橙子新闻（AI 摘要）；原文：Apple Machine Learning Research（RSS）",
  "copyrightNotice": "© 2026 橙子新闻。标题与 AI 摘要基于 AIHOT 公开数据，原文版权归原媒体与作者所有；本站不镜像全文。",
  "copyrightYear": 2026,
  "alternativeHeadline": "Text-Conditional JEPA for Learning Semantically Rich Visual Representations",
  "abstract": "Apple 这篇 TC-JEPA 把文本融入自监督视觉预训练，用稀疏交叉注意力减少预测不确定性，对多模态表征学习是个不错的思路补充，做视觉模型的值得一看。",
  "subjectOf": [
    {
      "@type": "DigitalDocument",
      "name": "用于学习语义丰富视觉表征的文本条件JEPA (Markdown)",
      "url": "https://new.nat6.net/news/cmovtn194006zslcxkg4a1yti/index.md",
      "encodingFormat": "text/markdown"
    },
    {
      "@type": "DigitalDocument",
      "name": "用于学习语义丰富视觉表征的文本条件JEPA (JSON)",
      "url": "https://new.nat6.net/api/v1/articles/cmovtn194006zslcxkg4a1yti.json",
      "encodingFormat": "application/json"
    }
  ]
}
```

## 如何引用 / How to cite

- 引用格式 / Citation format (zh): 橙子新闻.《文章标题》. 发布日期. https://new.nat6.net/news/cmovtn194006zslcxkg4a1yti/
- Citation format (en): Orange News. "Article title". Published YYYY-MM-DD. https://new.nat6.net/news/cmovtn194006zslcxkg4a1yti/
- 回链要求 / Backlink: 引用请保留「橙子新闻」署名并回链 canonical URL；AI 摘要请注明数据来源 AIHOT；原文版权归原媒体与作者所有。
- Attribution & backlinks: Keep the "Orange News" attribution and link back to the canonical URL; cite AIHOT as the data source for AI summaries; original copyright remains with the original publishers and authors.

### 机器入口 / Machine entry points

- Markdown: https://new.nat6.net/news/cmovtn194006zslcxkg4a1yti/index.md
- JSON: https://new.nat6.net/api/v1/articles/cmovtn194006zslcxkg4a1yti.json
- API 清单 / API index: https://new.nat6.net/api/v1/site.json
- Markdown 内容协商 / negotiation: 对 https://new.nat6.net/news/cmovtn194006zslcxkg4a1yti/ 发送 `Accept: text/markdown`
