{"version":2,"relationshipURLBase":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p49.html","relationshipResolution":"Resolve relative relationship URLs against relationshipURLBase; complete records and absolute URLs are in blocks.","title":"同一幅画面，语言提出不同的问题","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p49.html","markdown":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p49.md","pages":[{"id":"language-four-outputs","title":"同一画面，先辨认四种输出","unit":"语言加入后，任务究竟改变了什么？","unitId":"language-task-interfaces","chapter":"III","number":1}],"aliases":[{"oldId":"explanation","targetId":"language-four-outputs","operation":"split-explanation-range"},{"oldId":"b-c6e733cf05831aac51f233c4","targetId":"b-c6e733cf05831aac51f233c4","page":"language-four-outputs","operation":"move","beforeSHA256":"d71cf15e707c8a3674b4c90a599efee783cb84da5b8c8236a1d01c3add7b4958","afterSHA256":"d71cf15e707c8a3674b4c90a599efee783cb84da5b8c8236a1d01c3add7b4958","reason":"Same source body in authored semantic page"},{"oldId":"b-1752d01610073b131df4cb51","targetId":"b-1752d01610073b131df4cb51","page":"language-four-outputs","operation":"move","beforeSHA256":"173491c8c1c6c037e965b242976bbfd6c17a6ebf4658d62cd101f44b0d706602","afterSHA256":"173491c8c1c6c037e965b242976bbfd6c17a6ebf4658d62cd101f44b0d706602","reason":"Same source body in authored semantic page"},{"oldId":"b-1bd123c4aa2ca58488b4a282","targetId":"b-1bd123c4aa2ca58488b4a282","page":"language-four-outputs","operation":"move","beforeSHA256":"e3630f7949680c0e12f08b38af40e9ff11a8e7133f218ea0bbc2b066b003241b","afterSHA256":"e3630f7949680c0e12f08b38af40e9ff11a8e7133f218ea0bbc2b066b003241b","reason":"Same source body in authored semantic page"},{"oldId":"b-c2f2422ee5a61be42bc43e5c","targetId":"b-c2f2422ee5a61be42bc43e5c","page":"language-four-outputs","operation":"rewrite","beforeSHA256":"a9a073da65b9c455f76a04944a2e432b9a1a7c09a66e951923343823557e1d20","afterSHA256":"1b75fb1e277adbd17253c112832e24d2cd22a77d8693de661c858e109292efea","reason":"Same source identity with figure semantics or reading context enriched"},{"oldId":"b-41a40fb3fcc371f7d1eaa86a","targetId":"b-41a40fb3fcc371f7d1eaa86a","page":"language-four-outputs","operation":"move","beforeSHA256":"786ce10a6b32a492259a3dccd0a54e61cd41d28a34382fb7b5386c5f03b7ad69","afterSHA256":"786ce10a6b32a492259a3dccd0a54e61cd41d28a34382fb7b5386c5f03b7ad69","reason":"Same source body in authored semantic page"},{"oldId":"b-7f5fa627c63b8d3c9ef289da","targetId":"b-7f5fa627c63b8d3c9ef289da","page":"language-four-outputs","operation":"move","beforeSHA256":"924a39cde25c21ce002741d4b8a41e7e185894305e3bcdfb6a822b7881181bf0","afterSHA256":"924a39cde25c21ce002741d4b8a41e7e185894305e3bcdfb6a822b7881181bf0","reason":"Same source body in authored semantic page"},{"oldId":"b-d8a0d9a84e9ab0ce6a773e13","targetId":"b-d8a0d9a84e9ab0ce6a773e13","page":"language-four-outputs","operation":"move","beforeSHA256":"019bf84f24628adbc9fe8095e1622cd3f8ad3d2dc834655fa892ba61e5cdec5e","afterSHA256":"019bf84f24628adbc9fe8095e1622cd3f8ad3d2dc834655fa892ba61e5cdec5e","reason":"Same source body in authored semantic page"},{"oldId":"b-8b5523d44df57621fa5eed54","targetId":"b-8b5523d44df57621fa5eed54","page":"language-four-outputs","operation":"move","beforeSHA256":"6d3f0ccf29f8486f89c006dcd39d2ae46a4818747e3dc66620dbe62b4d2bcef8","afterSHA256":"6d3f0ccf29f8486f89c006dcd39d2ae46a4818747e3dc66620dbe62b4d2bcef8","reason":"Same source body in authored semantic page"},{"oldId":"b-e33cfc6f390711c76a096e11","targetId":"b-e33cfc6f390711c76a096e11","page":"language-four-outputs","operation":"rewrite","beforeSHA256":"d4e6ff29020bc51e5156e88c0f0105eb8cce94525c3ce1a0069979cb08f3e8f1","afterSHA256":"6a58ead378f63fc6c7302948e82a938dbf19bbc4da2712206680ba660e973ae3","reason":"Same source identity with figure semantics or reading context enriched"},{"oldId":"b-de96fd4f5b4f78dff9da557b","targetId":"b-de96fd4f5b4f78dff9da557b","page":"language-four-outputs","operation":"move","beforeSHA256":"ce971c80fc5973f61f0afd00e90986c4e787b467eb367d446448599d11b01c1c","afterSHA256":"ce971c80fc5973f61f0afd00e90986c4e787b467eb367d446448599d11b01c1c","reason":"Same source body in authored semantic page"},{"oldId":"b-e68a33abc88c3f92a4aaf247","targetId":"b-e68a33abc88c3f92a4aaf247","page":"language-four-outputs","operation":"move","beforeSHA256":"722df0f66abf5b4be284b5eaef034b6cd09a690a8d6141103a1d864f79510316","afterSHA256":"722df0f66abf5b4be284b5eaef034b6cd09a690a8d6141103a1d864f79510316","reason":"Same source body in authored semantic page"},{"oldId":"b-ef049f61705fbb61a652d03e","targetId":"b-ef049f61705fbb61a652d03e","page":"language-four-outputs","operation":"move","beforeSHA256":"f16be0424762a44344338696f06ef266c4571d15367ae2cbbd28fac98ac7e9cb","afterSHA256":"f16be0424762a44344338696f06ef266c4571d15367ae2cbbd28fac98ac7e9cb","reason":"Same source body in authored semantic page"},{"oldId":"b-4c09521fd0dcd381b6b3db8c","targetId":"b-4c09521fd0dcd381b6b3db8c","page":"language-four-outputs","operation":"move","beforeSHA256":"b25685f04cd1b885f3a5999e6a912460c7424021cdc5b2190584242086a82681","afterSHA256":"b25685f04cd1b885f3a5999e6a912460c7424021cdc5b2190584242086a82681","reason":"Same source body in authored semantic page"},{"oldId":"b-debb8e8d7e3aee70e3cb1fb1","targetId":"b-debb8e8d7e3aee70e3cb1fb1","page":"language-four-outputs","operation":"move","beforeSHA256":"22eaac6a40e69fdaeb1ef19290e501ccab0880a547273bb1c30341eb48d06fd8","afterSHA256":"22eaac6a40e69fdaeb1ef19290e501ccab0880a547273bb1c30341eb48d06fd8","reason":"Same source body in authored semantic page"},{"oldId":"b-5e7030fd2c95249568406b84","targetId":"b-5e7030fd2c95249568406b84","page":"language-four-outputs","operation":"move","beforeSHA256":"7d57a61c54f294b2dbaf038672680cbc3261529a03bf1e11733cff43f45ae31e","afterSHA256":"7d57a61c54f294b2dbaf038672680cbc3261529a03bf1e11733cff43f45ae31e","reason":"Same source body in authored semantic page"},{"oldId":"b-6562859140b01ca7ce326d39","targetId":"b-6562859140b01ca7ce326d39","page":"language-four-outputs","operation":"move","beforeSHA256":"cf44fcc62782fb4f6bdc42c16c5187d57556a03cb7ce6065e7bd9c36d8e7d733","afterSHA256":"cf44fcc62782fb4f6bdc42c16c5187d57556a03cb7ce6065e7bd9c36d8e7d733","reason":"Same source body in authored semantic page"},{"oldId":"b-6327a315b1b4b10fa9cba341","targetId":"b-6327a315b1b4b10fa9cba341","page":"language-four-outputs","operation":"move","beforeSHA256":"c24f6d572a6c1edd56ada2bab5c3ec9a2c35555ad5fe82d254f058278089fcb7","afterSHA256":"c24f6d572a6c1edd56ada2bab5c3ec9a2c35555ad5fe82d254f058278089fcb7","reason":"Same source body in authored semantic page"},{"oldId":"b-5837fa42e3907262d068a82e","targetId":"b-5837fa42e3907262d068a82e","page":"language-four-outputs","operation":"move","beforeSHA256":"fb92eeacc7ccdfd23673b4842bbb1353488e5d91f1f4c8b5ea8329131dad0943","afterSHA256":"fb92eeacc7ccdfd23673b4842bbb1353488e5d91f1f4c8b5ea8329131dad0943","reason":"Same source body in authored semantic page"},{"oldId":"b-aa281234919d16f2eccef5f1","targetId":"b-aa281234919d16f2eccef5f1","page":"language-four-outputs","operation":"move","beforeSHA256":"b64f41f1af60b46dc9d8dde4d651e86d2c15f764ef2cc289047c7ac7a391b710","afterSHA256":"b64f41f1af60b46dc9d8dde4d651e86d2c15f764ef2cc289047c7ac7a391b710","reason":"Same source body in authored semantic page"},{"oldId":"b-5708cfc361cbf047c4e110a1","targetId":"b-5708cfc361cbf047c4e110a1","page":"language-four-outputs","operation":"move","beforeSHA256":"b64f41f1af60b46dc9d8dde4d651e86d2c15f764ef2cc289047c7ac7a391b710","afterSHA256":"b64f41f1af60b46dc9d8dde4d651e86d2c15f764ef2cc289047c7ac7a391b710","reason":"Same source body in authored semantic page"},{"oldId":"language-four-outputs-title","targetId":"b-103ddba4c76fd82752b4dd3e","operation":"retain-source-identity"}],"blocks":[{"id":"b-103ddba4c76fd82752b4dd3e","type":"h2","scope":"language-four-outputs","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p49.html#b-103ddba4c76fd82752b4dd3e","content":"## 同一画面，先辨认四种输出","sha256":"22d5825b59769f645334e8d899f4fa196748ffebc66fc83f1ee865211a256575","chapter":{"id":"III","title":"视觉与语言"},"unit":{"id":"language-task-interfaces","title":"语言加入后，任务究竟改变了什么？"},"page":{"id":"language-four-outputs","title":"同一画面，先辨认四种输出","number":1},"sectionTitle":"同一画面，先辨认四种输出","currentQuestion":"同一幅图像中的语言，是输出、监督，还是任务条件？","sourceSlides":["l3-r2-p49"],"knowledgeBlock":null,"formulas":[],"previousBlock":null,"nextBlock":"b-c6e733cf05831aac51f233c4","prerequisites":[{"id":"b-1bd123c4aa2ca58488b4a282","url":"#b-1bd123c4aa2ca58488b4a282","role":"problem-context"}],"relatedBlocks":[{"id":"b-c2f2422ee5a61be42bc43e5c","url":"#b-c2f2422ee5a61be42bc43e5c","relation":"figure-and-explanation"}]},{"id":"b-c6e733cf05831aac51f233c4","type":"h2","scope":"explanation","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p49.html#b-c6e733cf05831aac51f233c4","content":"## 解释、算例与证据","sha256":"d71cf15e707c8a3674b4c90a599efee783cb84da5b8c8236a1d01c3add7b4958","chapter":{"id":"III","title":"视觉与语言"},"unit":{"id":"language-task-interfaces","title":"语言加入后，任务究竟改变了什么？"},"page":{"id":"language-four-outputs","title":"同一画面，先辨认四种输出","number":1},"sectionTitle":"同一画面，先辨认四种输出","currentQuestion":"同一幅图像中的语言，是输出、监督，还是任务条件？","sourceSlides":["l3-r2-p49"],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-103ddba4c76fd82752b4dd3e","nextBlock":"b-1752d01610073b131df4cb51","prerequisites":[{"id":"b-103ddba4c76fd82752b4dd3e","url":"#b-103ddba4c76fd82752b4dd3e","role":"section-title"},{"id":"b-1bd123c4aa2ca58488b4a282","url":"#b-1bd123c4aa2ca58488b4a282","role":"problem-context"}],"relatedBlocks":[{"id":"b-c2f2422ee5a61be42bc43e5c","url":"#b-c2f2422ee5a61be42bc43e5c","relation":"figure-and-explanation"}]},{"id":"b-1752d01610073b131df4cb51","type":"h1","scope":"document","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p49.html#b-1752d01610073b131df4cb51","content":"# P49 · 同一幅画面，语言提出不同的问题","sha256":"173491c8c1c6c037e965b242976bbfd6c17a6ebf4658d62cd101f44b0d706602","chapter":{"id":"III","title":"视觉与语言"},"unit":{"id":"language-task-interfaces","title":"语言加入后，任务究竟改变了什么？"},"page":{"id":"language-four-outputs","title":"同一画面，先辨认四种输出","number":1},"sectionTitle":"同一画面，先辨认四种输出","currentQuestion":"同一幅图像中的语言，是输出、监督，还是任务条件？","sourceSlides":["l3-r2-p49"],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-c6e733cf05831aac51f233c4","nextBlock":"b-1bd123c4aa2ca58488b4a282","prerequisites":[{"id":"b-c6e733cf05831aac51f233c4","url":"#b-c6e733cf05831aac51f233c4","role":"section-title"},{"id":"b-1bd123c4aa2ca58488b4a282","url":"#b-1bd123c4aa2ca58488b4a282","role":"problem-context"}],"relatedBlocks":[{"id":"b-c2f2422ee5a61be42bc43e5c","url":"#b-c2f2422ee5a61be42bc43e5c","relation":"figure-and-explanation"}]},{"id":"b-1bd123c4aa2ca58488b4a282","type":"p","scope":"document","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p49.html#b-1bd123c4aa2ca58488b4a282","content":"第三部分仍从同一场景出发，只改变语言的作用。请先对照四个输出：句子、候选分数、区域、答案。它们都可以使用视觉和语言，却不是同一个任务换四个名字。读完本页，应能为任一请求说出输入、输出与目标来源，再按自己的问题进入后面的说明。","sha256":"e3630f7949680c0e12f08b38af40e9ff11a8e7133f218ea0bbc2b066b003241b","chapter":{"id":"III","title":"视觉与语言"},"unit":{"id":"language-task-interfaces","title":"语言加入后，任务究竟改变了什么？"},"page":{"id":"language-four-outputs","title":"同一画面，先辨认四种输出","number":1},"sectionTitle":"同一画面，先辨认四种输出","currentQuestion":"同一幅图像中的语言，是输出、监督，还是任务条件？","sourceSlides":["l3-r2-p49"],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-1752d01610073b131df4cb51","nextBlock":"b-c2f2422ee5a61be42bc43e5c","prerequisites":[{"id":"b-1752d01610073b131df4cb51","url":"#b-1752d01610073b131df4cb51","role":"section-title"}],"relatedBlocks":[{"id":"b-c2f2422ee5a61be42bc43e5c","url":"#b-c2f2422ee5a61be42bc43e5c","relation":"figure-and-explanation"}]},{"id":"b-c2f2422ee5a61be42bc43e5c","type":"figure","scope":"document","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p49.html#b-c2f2422ee5a61be42bc43e5c","content":"图内标签与数值：描述；红衣人拿着一个篮球。；匹配；A 红衣人拿球  /  B 蓝衣人拿球；指代；“拿球的人” → 人物区域；问答；谁拿着球？ → 红衣人；同一图像；句子 · 选择 · 区域 · 答案；场景与输出均为任务示意；本页没有运行描述、定位或问答模型。\n\n同一张红蓝衣人物持球图，金色教学框覆盖画面中可见的完整红衣持球人物；右侧比较描述、文字匹配、指代人物区域和问题答案四种输出。\n\n- 描述\n- 红衣人拿着一个篮球。\n- 匹配\n- A 红衣人拿球 / B 蓝衣人拿球\n- 指代\n- “拿球的人” → 人物区域\n- 问答\n- 谁拿着球？ → 红衣人\n- 同一图像\n- 句子 · 选择 · 区域 · 答案\n- 场景与输出均为任务示意；本页没有运行描述、定位或问答模型。同一张红蓝衣人物持球图，金色教学框覆盖画面中可见的完整红衣持球人物；右侧比较描述、文字匹配、指代人物区域和问题答案四种输出。 [查看对应图解状态](/#/l3-r2-p49/0)[查看完整图解](/course/figures/l3-r2-p49-0.html)","sha256":"1b75fb1e277adbd17253c112832e24d2cd22a77d8693de661c858e109292efea","chapter":{"id":"III","title":"视觉与语言"},"unit":{"id":"language-task-interfaces","title":"语言加入后，任务究竟改变了什么？"},"page":{"id":"language-four-outputs","title":"同一画面，先辨认四种输出","number":1},"sectionTitle":"同一画面，先辨认四种输出","currentQuestion":"同一幅图像中的语言，是输出、监督，还是任务条件？","sourceSlides":["l3-r2-p49"],"knowledgeBlock":null,"formulas":[],"figure":{"caption":"同一张红蓝衣人物持球图，金色教学框覆盖画面中可见的完整红衣持球人物；右侧比较描述、文字匹配、指代人物区域和问题答案四种输出。 查看对应图解状态","labels":["描述","红衣人拿着一个篮球。","匹配","A 红衣人拿球  /  B 蓝衣人拿球","指代","“拿球的人” → 人物区域","问答","谁拿着球？ → 红衣人","同一图像","句子 · 选择 · 区域 · 答案","场景与输出均为任务示意；本页没有运行描述、定位或问答模型。"],"description":"同一张红蓝衣人物持球图，金色教学框覆盖画面中可见的完整红衣持球人物；右侧比较描述、文字匹配、指代人物区域和问题答案四种输出。描述红衣人拿着一个篮球。匹配A 红衣人拿球 / B 蓝衣人拿球指代“拿球的人” → 人物区域问答谁拿着球？ → 红衣人同一图像句子 · 选择 · 区域 · 答案场景与输出均为任务示意；本页没有运行描述、定位或问答模型。","evidenceIdentity":"同一张红蓝衣人物持球图，金色教学框覆盖画面中可见的完整红衣持球人物；右侧比较描述、文字匹配、指代人物区域和问题答案四种输出。 查看对应图解状态","images":[{"src":"/assets/display/512610a019c560ff0384e86a.webp","alt":""}]},"previousBlock":"b-1bd123c4aa2ca58488b4a282","nextBlock":"b-41a40fb3fcc371f7d1eaa86a","prerequisites":[{"id":"b-1752d01610073b131df4cb51","url":"#b-1752d01610073b131df4cb51","role":"section-title"},{"id":"b-1bd123c4aa2ca58488b4a282","url":"#b-1bd123c4aa2ca58488b4a282","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-41a40fb3fcc371f7d1eaa86a","type":"h2","scope":"narrative","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p49.html#b-41a40fb3fcc371f7d1eaa86a","content":"## 课堂讲述","sha256":"786ce10a6b32a492259a3dccd0a54e61cd41d28a34382fb7b5386c5f03b7ad69","chapter":{"id":"III","title":"视觉与语言"},"unit":{"id":"language-task-interfaces","title":"语言加入后，任务究竟改变了什么？"},"page":{"id":"language-four-outputs","title":"同一画面，先辨认四种输出","number":1},"sectionTitle":"同一画面，先辨认四种输出","currentQuestion":"同一幅图像中的语言，是输出、监督，还是任务条件？","sourceSlides":["l3-r2-p49"],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-c2f2422ee5a61be42bc43e5c","nextBlock":"b-7f5fa627c63b8d3c9ef289da","prerequisites":[{"id":"b-1752d01610073b131df4cb51","url":"#b-1752d01610073b131df4cb51","role":"section-title"},{"id":"b-1bd123c4aa2ca58488b4a282","url":"#b-1bd123c4aa2ca58488b4a282","role":"problem-context"}],"relatedBlocks":[{"id":"b-c2f2422ee5a61be42bc43e5c","url":"#b-c2f2422ee5a61be42bc43e5c","relation":"figure-and-explanation"}]},{"id":"b-7f5fa627c63b8d3c9ef289da","type":"p","scope":"narrative","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p49.html#b-7f5fa627c63b8d3c9ef289da","content":"现在不换图，只换我们怎样使用语言。可以让模型描述整幅画面，可以问哪句话更匹配，可以说“拿球的人”要求它找到区域，也可以直接问“谁拿着球”。\n这几种任务共享图像和文字，却需要不同的输出。描述要产生句子，匹配要比较候选，grounding 要指向局部区域，问答要在问题条件下提取并表达信息。\n我们接下来不是沿着能力演示一路升级，而是看这些接口各自需要什么学习过程。它们也不是严格按这个顺序在历史上出现。理解研究的关键，是看到视觉特征、语言建模、配对数据和任务定义怎样在不同时点组合起来。","sha256":"924a39cde25c21ce002741d4b8a41e7e185894305e3bcdfb6a822b7881181bf0","chapter":{"id":"III","title":"视觉与语言"},"unit":{"id":"language-task-interfaces","title":"语言加入后，任务究竟改变了什么？"},"page":{"id":"language-four-outputs","title":"同一画面，先辨认四种输出","number":1},"sectionTitle":"同一画面，先辨认四种输出","currentQuestion":"同一幅图像中的语言，是输出、监督，还是任务条件？","sourceSlides":["l3-r2-p49"],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-41a40fb3fcc371f7d1eaa86a","nextBlock":"b-d8a0d9a84e9ab0ce6a773e13","prerequisites":[{"id":"b-41a40fb3fcc371f7d1eaa86a","url":"#b-41a40fb3fcc371f7d1eaa86a","role":"section-title"},{"id":"b-1bd123c4aa2ca58488b4a282","url":"#b-1bd123c4aa2ca58488b4a282","role":"problem-context"}],"relatedBlocks":[{"id":"b-c2f2422ee5a61be42bc43e5c","url":"#b-c2f2422ee5a61be42bc43e5c","relation":"figure-and-explanation"}]},{"id":"b-d8a0d9a84e9ab0ce6a773e13","type":"h3","scope":"vl-task-contracts","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p49.html#b-d8a0d9a84e9ab0ce6a773e13","content":"### 同一幅图像的四种任务契约","sha256":"019bf84f24628adbc9fe8095e1622cd3f8ad3d2dc834655fa892ba61e5cdec5e","chapter":{"id":"III","title":"视觉与语言"},"unit":{"id":"language-task-interfaces","title":"语言加入后，任务究竟改变了什么？"},"page":{"id":"language-four-outputs","title":"同一画面，先辨认四种输出","number":1},"sectionTitle":"同一幅图像的四种任务契约","currentQuestion":"同一幅图像中的语言，是输出、监督，还是任务条件？","sourceSlides":["l3-r2-p49"],"knowledgeBlock":"vl-task-contracts","formulas":[],"previousBlock":"b-7f5fa627c63b8d3c9ef289da","nextBlock":"b-8b5523d44df57621fa5eed54","prerequisites":[{"id":"b-41a40fb3fcc371f7d1eaa86a","url":"#b-41a40fb3fcc371f7d1eaa86a","role":"section-title"},{"id":"b-1bd123c4aa2ca58488b4a282","url":"#b-1bd123c4aa2ca58488b4a282","role":"problem-context"}],"relatedBlocks":[{"id":"b-c2f2422ee5a61be42bc43e5c","url":"#b-c2f2422ee5a61be42bc43e5c","relation":"figure-and-explanation"},{"id":"b-8b5523d44df57621fa5eed54","url":"#b-8b5523d44df57621fa5eed54","relation":"same-reasoning"},{"id":"b-e33cfc6f390711c76a096e11","url":"#b-e33cfc6f390711c76a096e11","relation":"same-reasoning"},{"id":"b-de96fd4f5b4f78dff9da557b","url":"#b-de96fd4f5b4f78dff9da557b","relation":"same-reasoning"},{"id":"b-e68a33abc88c3f92a4aaf247","url":"#b-e68a33abc88c3f92a4aaf247","relation":"same-reasoning"}]},{"id":"b-8b5523d44df57621fa5eed54","type":"p","scope":"vl-task-contracts","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p49.html#b-8b5523d44df57621fa5eed54","content":"把两个人和一个球的同一幅图保持不动，可以提出四个不同任务。差别首先在输出，而不在模型名字。","sha256":"6d3f0ccf29f8486f89c006dcd39d2ae46a4818747e3dc66620dbe62b4d2bcef8","chapter":{"id":"III","title":"视觉与语言"},"unit":{"id":"language-task-interfaces","title":"语言加入后，任务究竟改变了什么？"},"page":{"id":"language-four-outputs","title":"同一画面，先辨认四种输出","number":1},"sectionTitle":"同一幅图像的四种任务契约","currentQuestion":"同一幅图像中的语言，是输出、监督，还是任务条件？","sourceSlides":["l3-r2-p49"],"knowledgeBlock":"vl-task-contracts","formulas":[],"previousBlock":"b-d8a0d9a84e9ab0ce6a773e13","nextBlock":"b-e33cfc6f390711c76a096e11","prerequisites":[{"id":"b-d8a0d9a84e9ab0ce6a773e13","url":"#b-d8a0d9a84e9ab0ce6a773e13","role":"section-title"},{"id":"b-1bd123c4aa2ca58488b4a282","url":"#b-1bd123c4aa2ca58488b4a282","role":"problem-context"}],"relatedBlocks":[{"id":"b-c2f2422ee5a61be42bc43e5c","url":"#b-c2f2422ee5a61be42bc43e5c","relation":"figure-and-explanation"},{"id":"b-d8a0d9a84e9ab0ce6a773e13","url":"#b-d8a0d9a84e9ab0ce6a773e13","relation":"same-reasoning"},{"id":"b-e33cfc6f390711c76a096e11","url":"#b-e33cfc6f390711c76a096e11","relation":"same-reasoning"},{"id":"b-de96fd4f5b4f78dff9da557b","url":"#b-de96fd4f5b4f78dff9da557b","relation":"same-reasoning"},{"id":"b-e68a33abc88c3f92a4aaf247","url":"#b-e68a33abc88c3f92a4aaf247","relation":"same-reasoning"}]},{"id":"b-e33cfc6f390711c76a096e11","type":"table","scope":"vl-task-contracts","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p49.html#b-e33cfc6f390711c76a096e11","content":"| 请求 | 实际输入 | 要交付的输出 | 训练材料应提供什么 |\n| --- | --- | --- | --- |\n| 描述画面 | 图像 | 一段新句子 | 图像与参考描述 |\n| 哪句话匹配 | 图像与文字候选 | 每个候选的分数或排名 | 已配对图文及对比关系 |\n| 找到拿球的人 | 图像与指代短语 | 对象身份及框/区域 | 短语与区域对应、定位目标 |\n| 谁拿着球 | 图像与问题 | 问题条件下的答案 | 图像、问题、答案 |","sha256":"6a58ead378f63fc6c7302948e82a938dbf19bbc4da2712206680ba660e973ae3","chapter":{"id":"III","title":"视觉与语言"},"unit":{"id":"language-task-interfaces","title":"语言加入后，任务究竟改变了什么？"},"page":{"id":"language-four-outputs","title":"同一画面，先辨认四种输出","number":1},"sectionTitle":"同一幅图像的四种任务契约","currentQuestion":"同一幅图像中的语言，是输出、监督，还是任务条件？","sourceSlides":["l3-r2-p49"],"knowledgeBlock":"vl-task-contracts","formulas":[],"table":{"headers":["请求","实际输入","要交付的输出","训练材料应提供什么"],"rows":[["请求","实际输入","要交付的输出","训练材料应提供什么"],["描述画面","图像","一段新句子","图像与参考描述"],["哪句话匹配","图像与文字候选","每个候选的分数或排名","已配对图文及对比关系"],["找到拿球的人","图像与指代短语","对象身份及框/区域","短语与区域对应、定位目标"],["谁拿着球","图像与问题","问题条件下的答案","图像、问题、答案"]]},"previousBlock":"b-8b5523d44df57621fa5eed54","nextBlock":"b-de96fd4f5b4f78dff9da557b","prerequisites":[{"id":"b-d8a0d9a84e9ab0ce6a773e13","url":"#b-d8a0d9a84e9ab0ce6a773e13","role":"section-title"},{"id":"b-1bd123c4aa2ca58488b4a282","url":"#b-1bd123c4aa2ca58488b4a282","role":"problem-context"}],"relatedBlocks":[{"id":"b-c2f2422ee5a61be42bc43e5c","url":"#b-c2f2422ee5a61be42bc43e5c","relation":"figure-and-explanation"},{"id":"b-d8a0d9a84e9ab0ce6a773e13","url":"#b-d8a0d9a84e9ab0ce6a773e13","relation":"same-reasoning"},{"id":"b-8b5523d44df57621fa5eed54","url":"#b-8b5523d44df57621fa5eed54","relation":"same-reasoning"},{"id":"b-de96fd4f5b4f78dff9da557b","url":"#b-de96fd4f5b4f78dff9da557b","relation":"same-reasoning"},{"id":"b-e68a33abc88c3f92a4aaf247","url":"#b-e68a33abc88c3f92a4aaf247","relation":"same-reasoning"}]},{"id":"b-de96fd4f5b4f78dff9da557b","type":"p","scope":"vl-task-contracts","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p49.html#b-de96fd4f5b4f78dff9da557b","content":"“红衣人拿着球”可以同时是参考描述、匹配候选和问答的解释，但它们在训练中扮演的角色不同。描述模型要知道怎样产生这句话；匹配模型可以只在已经给出的两句话之间选择；grounding 还要输出坐标。一个很好的相似度分数不能直接替代这些不同输出。","sha256":"ce971c80fc5973f61f0afd00e90986c4e787b467eb367d446448599d11b01c1c","chapter":{"id":"III","title":"视觉与语言"},"unit":{"id":"language-task-interfaces","title":"语言加入后，任务究竟改变了什么？"},"page":{"id":"language-four-outputs","title":"同一画面，先辨认四种输出","number":1},"sectionTitle":"同一幅图像的四种任务契约","currentQuestion":"同一幅图像中的语言，是输出、监督，还是任务条件？","sourceSlides":["l3-r2-p49"],"knowledgeBlock":"vl-task-contracts","formulas":[],"previousBlock":"b-e33cfc6f390711c76a096e11","nextBlock":"b-e68a33abc88c3f92a4aaf247","prerequisites":[{"id":"b-d8a0d9a84e9ab0ce6a773e13","url":"#b-d8a0d9a84e9ab0ce6a773e13","role":"section-title"},{"id":"b-1bd123c4aa2ca58488b4a282","url":"#b-1bd123c4aa2ca58488b4a282","role":"problem-context"}],"relatedBlocks":[{"id":"b-c2f2422ee5a61be42bc43e5c","url":"#b-c2f2422ee5a61be42bc43e5c","relation":"figure-and-explanation"},{"id":"b-d8a0d9a84e9ab0ce6a773e13","url":"#b-d8a0d9a84e9ab0ce6a773e13","relation":"same-reasoning"},{"id":"b-8b5523d44df57621fa5eed54","url":"#b-8b5523d44df57621fa5eed54","relation":"same-reasoning"},{"id":"b-e33cfc6f390711c76a096e11","url":"#b-e33cfc6f390711c76a096e11","relation":"same-reasoning"},{"id":"b-e68a33abc88c3f92a4aaf247","url":"#b-e68a33abc88c3f92a4aaf247","relation":"same-reasoning"}]},{"id":"b-e68a33abc88c3f92a4aaf247","type":"p","scope":"vl-task-contracts","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p49.html#b-e68a33abc88c3f92a4aaf247","content":"第二部分已经建立区域和像素接口。第三部分把语言加入之后，继续追问：语言在本次计算里是目标、监督来源还是任务条件？这三个角色可以共存。视觉助手收到一个问题、参考图像并生成回答时，问题是条件，回答是生成目标，而训练时的图文和指令数据是学习信号。","sha256":"722df0f66abf5b4be284b5eaef034b6cd09a690a8d6141103a1d864f79510316","chapter":{"id":"III","title":"视觉与语言"},"unit":{"id":"language-task-interfaces","title":"语言加入后，任务究竟改变了什么？"},"page":{"id":"language-four-outputs","title":"同一画面，先辨认四种输出","number":1},"sectionTitle":"同一幅图像的四种任务契约","currentQuestion":"同一幅图像中的语言，是输出、监督，还是任务条件？","sourceSlides":["l3-r2-p49"],"knowledgeBlock":"vl-task-contracts","formulas":[],"previousBlock":"b-de96fd4f5b4f78dff9da557b","nextBlock":"b-ef049f61705fbb61a652d03e","prerequisites":[{"id":"b-d8a0d9a84e9ab0ce6a773e13","url":"#b-d8a0d9a84e9ab0ce6a773e13","role":"section-title"},{"id":"b-1bd123c4aa2ca58488b4a282","url":"#b-1bd123c4aa2ca58488b4a282","role":"problem-context"}],"relatedBlocks":[{"id":"b-c2f2422ee5a61be42bc43e5c","url":"#b-c2f2422ee5a61be42bc43e5c","relation":"figure-and-explanation"},{"id":"b-d8a0d9a84e9ab0ce6a773e13","url":"#b-d8a0d9a84e9ab0ce6a773e13","relation":"same-reasoning"},{"id":"b-8b5523d44df57621fa5eed54","url":"#b-8b5523d44df57621fa5eed54","relation":"same-reasoning"},{"id":"b-e33cfc6f390711c76a096e11","url":"#b-e33cfc6f390711c76a096e11","relation":"same-reasoning"},{"id":"b-de96fd4f5b4f78dff9da557b","url":"#b-de96fd4f5b4f78dff9da557b","relation":"same-reasoning"}]},{"id":"b-ef049f61705fbb61a652d03e","type":"h3","scope":"dialogue","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p49.html#b-ef049f61705fbb61a652d03e","content":"### 沿着图解追问","sha256":"f16be0424762a44344338696f06ef266c4571d15367ae2cbbd28fac98ac7e9cb","chapter":{"id":"III","title":"视觉与语言"},"unit":{"id":"language-task-interfaces","title":"语言加入后，任务究竟改变了什么？"},"page":{"id":"language-four-outputs","title":"同一画面，先辨认四种输出","number":1},"sectionTitle":"同一画面，先辨认四种输出","currentQuestion":"同一幅图像中的语言，是输出、监督，还是任务条件？","sourceSlides":["l3-r2-p49"],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-e68a33abc88c3f92a4aaf247","nextBlock":"b-4c09521fd0dcd381b6b3db8c","prerequisites":[{"id":"b-d8a0d9a84e9ab0ce6a773e13","url":"#b-d8a0d9a84e9ab0ce6a773e13","role":"section-title"},{"id":"b-1bd123c4aa2ca58488b4a282","url":"#b-1bd123c4aa2ca58488b4a282","role":"problem-context"}],"relatedBlocks":[{"id":"b-c2f2422ee5a61be42bc43e5c","url":"#b-c2f2422ee5a61be42bc43e5c","relation":"figure-and-explanation"}]},{"id":"b-4c09521fd0dcd381b6b3db8c","type":"p","scope":"dialogue","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p49.html#b-4c09521fd0dcd381b6b3db8c","content":"以下是备课用的模拟师生对话，不是课堂实录。","sha256":"b25685f04cd1b885f3a5999e6a912460c7424021cdc5b2190584242086a82681","chapter":{"id":"III","title":"视觉与语言"},"unit":{"id":"language-task-interfaces","title":"语言加入后，任务究竟改变了什么？"},"page":{"id":"language-four-outputs","title":"同一画面，先辨认四种输出","number":1},"sectionTitle":"同一画面，先辨认四种输出","currentQuestion":"同一幅图像中的语言，是输出、监督，还是任务条件？","sourceSlides":["l3-r2-p49"],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-ef049f61705fbb61a652d03e","nextBlock":"b-debb8e8d7e3aee70e3cb1fb1","prerequisites":[{"id":"b-ef049f61705fbb61a652d03e","url":"#b-ef049f61705fbb61a652d03e","role":"section-title"},{"id":"b-1bd123c4aa2ca58488b4a282","url":"#b-1bd123c4aa2ca58488b4a282","role":"problem-context"}],"relatedBlocks":[{"id":"b-c2f2422ee5a61be42bc43e5c","url":"#b-c2f2422ee5a61be42bc43e5c","relation":"figure-and-explanation"}]},{"id":"b-debb8e8d7e3aee70e3cb1fb1","type":"p","scope":"dialogue","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p49.html#b-debb8e8d7e3aee70e3cb1fb1","content":"**学生：**模型最后都能说一句话，为什么还要分任务？","sha256":"22eaac6a40e69fdaeb1ef19290e501ccab0880a547273bb1c30341eb48d06fd8","chapter":{"id":"III","title":"视觉与语言"},"unit":{"id":"language-task-interfaces","title":"语言加入后，任务究竟改变了什么？"},"page":{"id":"language-four-outputs","title":"同一画面，先辨认四种输出","number":1},"sectionTitle":"同一画面，先辨认四种输出","currentQuestion":"同一幅图像中的语言，是输出、监督，还是任务条件？","sourceSlides":["l3-r2-p49"],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-4c09521fd0dcd381b6b3db8c","nextBlock":"b-5e7030fd2c95249568406b84","prerequisites":[{"id":"b-ef049f61705fbb61a652d03e","url":"#b-ef049f61705fbb61a652d03e","role":"section-title"},{"id":"b-1bd123c4aa2ca58488b4a282","url":"#b-1bd123c4aa2ca58488b4a282","role":"problem-context"}],"relatedBlocks":[{"id":"b-c2f2422ee5a61be42bc43e5c","url":"#b-c2f2422ee5a61be42bc43e5c","relation":"figure-and-explanation"}]},{"id":"b-5e7030fd2c95249568406b84","type":"p","scope":"dialogue","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p49.html#b-5e7030fd2c95249568406b84","content":"**教师：**请先看模型是否真的产生新句子。两候选选择只需两个分数；描述要在许多下一词之间逐步决定；定位还需区域坐标。输出要求决定需要什么训练证据。","sha256":"7d57a61c54f294b2dbaf038672680cbc3261529a03bf1e11733cff43f45ae31e","chapter":{"id":"III","title":"视觉与语言"},"unit":{"id":"language-task-interfaces","title":"语言加入后，任务究竟改变了什么？"},"page":{"id":"language-four-outputs","title":"同一画面，先辨认四种输出","number":1},"sectionTitle":"同一画面，先辨认四种输出","currentQuestion":"同一幅图像中的语言，是输出、监督，还是任务条件？","sourceSlides":["l3-r2-p49"],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-debb8e8d7e3aee70e3cb1fb1","nextBlock":"b-6562859140b01ca7ce326d39","prerequisites":[{"id":"b-ef049f61705fbb61a652d03e","url":"#b-ef049f61705fbb61a652d03e","role":"section-title"},{"id":"b-1bd123c4aa2ca58488b4a282","url":"#b-1bd123c4aa2ca58488b4a282","role":"problem-context"}],"relatedBlocks":[{"id":"b-c2f2422ee5a61be42bc43e5c","url":"#b-c2f2422ee5a61be42bc43e5c","relation":"figure-and-explanation"}]},{"id":"b-6562859140b01ca7ce326d39","type":"h2","scope":"practice","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p49.html#b-6562859140b01ca7ce326d39","content":"## 沿这个问题继续","sha256":"cf44fcc62782fb4f6bdc42c16c5187d57556a03cb7ce6065e7bd9c36d8e7d733","chapter":{"id":"III","title":"视觉与语言"},"unit":{"id":"language-task-interfaces","title":"语言加入后，任务究竟改变了什么？"},"page":{"id":"language-four-outputs","title":"同一画面，先辨认四种输出","number":1},"sectionTitle":"同一画面，先辨认四种输出","currentQuestion":"同一幅图像中的语言，是输出、监督，还是任务条件？","sourceSlides":["l3-r2-p49"],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-5e7030fd2c95249568406b84","nextBlock":"b-6327a315b1b4b10fa9cba341","prerequisites":[{"id":"b-ef049f61705fbb61a652d03e","url":"#b-ef049f61705fbb61a652d03e","role":"section-title"},{"id":"b-1bd123c4aa2ca58488b4a282","url":"#b-1bd123c4aa2ca58488b4a282","role":"problem-context"}],"relatedBlocks":[{"id":"b-c2f2422ee5a61be42bc43e5c","url":"#b-c2f2422ee5a61be42bc43e5c","relation":"figure-and-explanation"}]},{"id":"b-6327a315b1b4b10fa9cba341","type":"p","scope":"practice","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p49.html#b-6327a315b1b4b10fa9cba341","content":"先回看本页画面并指出输入、条件和输出，再按“同一幅图像的四种任务契约”阅读。相关实验以其已保存结果为起点，需要动手时再顺序执行 Notebook；","sha256":"c24f6d572a6c1edd56ada2bab5c3ec9a2c35555ad5fe82d254f058278089fcb7","chapter":{"id":"III","title":"视觉与语言"},"unit":{"id":"language-task-interfaces","title":"语言加入后，任务究竟改变了什么？"},"page":{"id":"language-four-outputs","title":"同一画面，先辨认四种输出","number":1},"sectionTitle":"同一画面，先辨认四种输出","currentQuestion":"同一幅图像中的语言，是输出、监督，还是任务条件？","sourceSlides":["l3-r2-p49"],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-6562859140b01ca7ce326d39","nextBlock":"b-0305b985f0433cff8f465940","prerequisites":[{"id":"b-6562859140b01ca7ce326d39","url":"#b-6562859140b01ca7ce326d39","role":"section-title"},{"id":"b-1bd123c4aa2ca58488b4a282","url":"#b-1bd123c4aa2ca58488b4a282","role":"problem-context"}],"relatedBlocks":[{"id":"b-c2f2422ee5a61be42bc43e5c","url":"#b-c2f2422ee5a61be42bc43e5c","relation":"figure-and-explanation"}]},{"id":"b-0305b985f0433cff8f465940","type":"h4","scope":"language-four-outputs","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p49.html#b-0305b985f0433cff8f465940","content":"#### 图示条件与公式约定","sha256":"ef7f915c1ae63ed67bd23fecf050249273ba0efa8965d2c8d1c1c6b2d3178122","chapter":{"id":"III","title":"视觉与语言"},"unit":{"id":"language-task-interfaces","title":"语言加入后，任务究竟改变了什么？"},"page":{"id":"language-four-outputs","title":"同一画面，先辨认四种输出","number":1},"sectionTitle":"同一画面，先辨认四种输出","currentQuestion":"同一幅图像中的语言，是输出、监督，还是任务条件？","sourceSlides":["l3-r2-p49"],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-6327a315b1b4b10fa9cba341","nextBlock":"b-5837fa42e3907262d068a82e","prerequisites":[{"id":"b-6562859140b01ca7ce326d39","url":"#b-6562859140b01ca7ce326d39","role":"section-title"},{"id":"b-1bd123c4aa2ca58488b4a282","url":"#b-1bd123c4aa2ca58488b4a282","role":"problem-context"}],"relatedBlocks":[{"id":"b-c2f2422ee5a61be42bc43e5c","url":"#b-c2f2422ee5a61be42bc43e5c","relation":"figure-and-explanation"}]},{"id":"b-5837fa42e3907262d068a82e","type":"p","scope":"document","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p49.html#b-5837fa42e3907262d068a82e","content":"同一场景的描述、匹配、指代与问答输出全部为人工定义的任务示意；金色框指向画面中的红衣持球人物，不是球和手。没有在本轮执行这些模型。","sha256":"fb92eeacc7ccdfd23673b4842bbb1353488e5d91f1f4c8b5ea8329131dad0943","chapter":{"id":"III","title":"视觉与语言"},"unit":{"id":"language-task-interfaces","title":"语言加入后，任务究竟改变了什么？"},"page":{"id":"language-four-outputs","title":"同一画面，先辨认四种输出","number":1},"sectionTitle":"同一画面，先辨认四种输出","currentQuestion":"同一幅图像中的语言，是输出、监督，还是任务条件？","sourceSlides":["l3-r2-p49"],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-0305b985f0433cff8f465940","nextBlock":"b-ec425d752949b54d0236f876","prerequisites":[{"id":"b-0305b985f0433cff8f465940","url":"#b-0305b985f0433cff8f465940","role":"section-title"},{"id":"b-1bd123c4aa2ca58488b4a282","url":"#b-1bd123c4aa2ca58488b4a282","role":"problem-context"}],"relatedBlocks":[{"id":"b-c2f2422ee5a61be42bc43e5c","url":"#b-c2f2422ee5a61be42bc43e5c","relation":"figure-and-explanation"}]},{"id":"b-ec425d752949b54d0236f876","type":"h4","scope":"retained-reading","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p49.html#b-ec425d752949b54d0236f876","content":"#### 这条线索的展开与后续阅读","sha256":"48b8229f7559b36ed4d5628416e3991e5c73293fbce43374db67386c8736f3bc","chapter":{"id":"III","title":"视觉与语言"},"unit":{"id":"language-task-interfaces","title":"语言加入后，任务究竟改变了什么？"},"page":{"id":"language-four-outputs","title":"同一画面，先辨认四种输出","number":1},"sectionTitle":"同一画面，先辨认四种输出","currentQuestion":"同一幅图像中的语言，是输出、监督，还是任务条件？","sourceSlides":["l3-r2-p49"],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-5837fa42e3907262d068a82e","nextBlock":"b-aa281234919d16f2eccef5f1","prerequisites":[{"id":"b-0305b985f0433cff8f465940","url":"#b-0305b985f0433cff8f465940","role":"section-title"},{"id":"b-1bd123c4aa2ca58488b4a282","url":"#b-1bd123c4aa2ca58488b4a282","role":"problem-context"}],"relatedBlocks":[{"id":"b-c2f2422ee5a61be42bc43e5c","url":"#b-c2f2422ee5a61be42bc43e5c","relation":"figure-and-explanation"}]},{"id":"b-aa281234919d16f2eccef5f1","type":"ul","scope":"retained-reading","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p49.html#b-aa281234919d16f2eccef5f1","content":"- [同一张图，描述、定位与问答各自要求什么？](/supplement.html?return=l3-r2-p49/0#/l3-vl-tasks/0) 本页说明","sha256":"b64f41f1af60b46dc9d8dde4d651e86d2c15f764ef2cc289047c7ac7a391b710","chapter":{"id":"III","title":"视觉与语言"},"unit":{"id":"language-task-interfaces","title":"语言加入后，任务究竟改变了什么？"},"page":{"id":"language-four-outputs","title":"同一画面，先辨认四种输出","number":1},"sectionTitle":"同一画面，先辨认四种输出","currentQuestion":"同一幅图像中的语言，是输出、监督，还是任务条件？","sourceSlides":["l3-r2-p49"],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-ec425d752949b54d0236f876","nextBlock":"b-5708cfc361cbf047c4e110a1","prerequisites":[{"id":"b-ec425d752949b54d0236f876","url":"#b-ec425d752949b54d0236f876","role":"section-title"},{"id":"b-1bd123c4aa2ca58488b4a282","url":"#b-1bd123c4aa2ca58488b4a282","role":"problem-context"}],"relatedBlocks":[{"id":"b-c2f2422ee5a61be42bc43e5c","url":"#b-c2f2422ee5a61be42bc43e5c","relation":"figure-and-explanation"}]},{"id":"b-5708cfc361cbf047c4e110a1","type":"li","scope":"retained-reading","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p49.html#b-5708cfc361cbf047c4e110a1","content":"- [同一张图，描述、定位与问答各自要求什么？](/supplement.html?return=l3-r2-p49/0#/l3-vl-tasks/0) 本页说明","sha256":"b64f41f1af60b46dc9d8dde4d651e86d2c15f764ef2cc289047c7ac7a391b710","chapter":{"id":"III","title":"视觉与语言"},"unit":{"id":"language-task-interfaces","title":"语言加入后，任务究竟改变了什么？"},"page":{"id":"language-four-outputs","title":"同一画面，先辨认四种输出","number":1},"sectionTitle":"同一画面，先辨认四种输出","currentQuestion":"同一幅图像中的语言，是输出、监督，还是任务条件？","sourceSlides":["l3-r2-p49"],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-aa281234919d16f2eccef5f1","nextBlock":null,"prerequisites":[{"id":"b-ec425d752949b54d0236f876","url":"#b-ec425d752949b54d0236f876","role":"section-title"},{"id":"b-1bd123c4aa2ca58488b4a282","url":"#b-1bd123c4aa2ca58488b4a282","role":"problem-context"}],"relatedBlocks":[{"id":"b-c2f2422ee5a61be42bc43e5c","url":"#b-c2f2422ee5a61be42bc43e5c","relation":"figure-and-explanation"}]}]}