{"version":2,"relationshipURLBase":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p44.html","relationshipResolution":"Resolve relative relationship URLs against relationshipURLBase; complete records and absolute URLs are in blocks.","title":"两种布局，如果已经变成同一个数","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p44.html","markdown":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p44.md","pages":[{"id":"representation-identical-inputs","title":"固定相同表示，推导为什么无法区分左右","unit":"瓶颈是读错位置，还是差别根本没有保留？","unitId":"spatial-information-preservation","chapter":"II","number":1},{"id":"representation-reading-diagnosis","title":"把对齐收益与信息限制放回各自条件","unit":"瓶颈是读错位置，还是差别根本没有保留？","unitId":"spatial-information-preservation","chapter":"II","number":2}],"aliases":[{"oldId":"explanation","targetId":"representation-identical-inputs","operation":"split-explanation-range"},{"oldId":"b-5352ab09b69fecb5bf8bc01a","targetId":"b-5352ab09b69fecb5bf8bc01a","page":"representation-identical-inputs","operation":"move","beforeSHA256":"d71cf15e707c8a3674b4c90a599efee783cb84da5b8c8236a1d01c3add7b4958","afterSHA256":"d71cf15e707c8a3674b4c90a599efee783cb84da5b8c8236a1d01c3add7b4958","reason":"Same source body in authored semantic page"},{"oldId":"b-de79c86b47d5f73e04c8d7aa","targetId":"b-de79c86b47d5f73e04c8d7aa","page":"representation-identical-inputs","operation":"move","beforeSHA256":"f4ec8b93ef44d945f955f22f4af5769481fd54084b82e0f15b7101aa9488a466","afterSHA256":"f4ec8b93ef44d945f955f22f4af5769481fd54084b82e0f15b7101aa9488a466","reason":"Same source body in authored semantic page"},{"oldId":"b-915b81c2c7d0b70935bc5355","targetId":"b-915b81c2c7d0b70935bc5355","page":"representation-identical-inputs","operation":"move","beforeSHA256":"bb805a18dc186d875e39921d6b91f9d80a17f61043f941f55e961c2edc4fad3e","afterSHA256":"bb805a18dc186d875e39921d6b91f9d80a17f61043f941f55e961c2edc4fad3e","reason":"Same source body in authored semantic page"},{"oldId":"b-d095f89c111a019d699824fd","targetId":"b-d095f89c111a019d699824fd","page":"representation-identical-inputs","operation":"rewrite","beforeSHA256":"970c97fbe436ff1715f682960325b2a63ae18bc546594222772143be4d8a7240","afterSHA256":"9a8dad14132213d371a29c72d35df52c915bb4e750b83704541f5e1846a4eed2","reason":"Same source identity with figure semantics or reading context enriched"},{"oldId":"b-09d657eb109005c80d10213b","targetId":"b-09d657eb109005c80d10213b","page":"representation-identical-inputs","operation":"move","beforeSHA256":"786ce10a6b32a492259a3dccd0a54e61cd41d28a34382fb7b5386c5f03b7ad69","afterSHA256":"786ce10a6b32a492259a3dccd0a54e61cd41d28a34382fb7b5386c5f03b7ad69","reason":"Same source body in authored semantic page"},{"oldId":"b-b9ce927f3e29d58d7ef5e59a","targetId":"b-b9ce927f3e29d58d7ef5e59a","page":"representation-identical-inputs","operation":"move","beforeSHA256":"7a149385c08b9a3671905a2798fb27bb47f83bba2b6d0b3940a4c3425c2117df","afterSHA256":"7a149385c08b9a3671905a2798fb27bb47f83bba2b6d0b3940a4c3425c2117df","reason":"Same source body in authored semantic page"},{"oldId":"b-7d77ceff9eb966c8abc9598e","targetId":"b-7d77ceff9eb966c8abc9598e","page":"representation-identical-inputs","operation":"move","beforeSHA256":"b07830a9aa685e99757cc9b85c7a56ca2ea992afef9d9b9f8e88c83e45b6a619","afterSHA256":"b07830a9aa685e99757cc9b85c7a56ca2ea992afef9d9b9f8e88c83e45b6a619","reason":"Same source body in authored semantic page"},{"oldId":"b-31f7116af39dca2901da23fe","targetId":"b-31f7116af39dca2901da23fe","page":"representation-identical-inputs","operation":"move","beforeSHA256":"bd243b07b2f5a86fb352c34ed9d40f637db6e093f7c59e8a243871e20a22699a","afterSHA256":"bd243b07b2f5a86fb352c34ed9d40f637db6e093f7c59e8a243871e20a22699a","reason":"Same source body in authored semantic page"},{"oldId":"b-26caeac455cc1c2c6464ebf7","targetId":"b-26caeac455cc1c2c6464ebf7","page":"representation-identical-inputs","operation":"move","beforeSHA256":"e77e8ad77aa56049b6fb8247fda69458bb1544e738be5d381781019ffb534dd2","afterSHA256":"e77e8ad77aa56049b6fb8247fda69458bb1544e738be5d381781019ffb534dd2","reason":"Same source body in authored semantic page"},{"oldId":"b-a9a56bfbc46c888f0d322697","targetId":"b-a9a56bfbc46c888f0d322697","page":"representation-identical-inputs","operation":"move","beforeSHA256":"3d80b9ffcb056ac171153d7e1ef1eb3ad3c5e3d63645eab0b9ddd628ac06925a","afterSHA256":"3d80b9ffcb056ac171153d7e1ef1eb3ad3c5e3d63645eab0b9ddd628ac06925a","reason":"Same source body in authored semantic page"},{"oldId":"b-9cd63cc5f0f05ca3332d6784","targetId":"b-9cd63cc5f0f05ca3332d6784","page":"representation-identical-inputs","operation":"move","beforeSHA256":"3de609c492ec57041f21576d7d8f5f74cfe093c085b75f1ea22b5d56980ed070","afterSHA256":"3de609c492ec57041f21576d7d8f5f74cfe093c085b75f1ea22b5d56980ed070","reason":"Same source body in authored semantic page"},{"oldId":"b-5a237dc21ad2fda4ff2cf1db","targetId":"b-5a237dc21ad2fda4ff2cf1db","page":"representation-identical-inputs","operation":"move","beforeSHA256":"59c8e843f3c5c93621033a4af327bb27e577a998f65f8b32398837089055ea5c","afterSHA256":"59c8e843f3c5c93621033a4af327bb27e577a998f65f8b32398837089055ea5c","reason":"Same source body in authored semantic page"},{"oldId":"b-0cc7a89d266bb20a7ff018c0","targetId":"b-0cc7a89d266bb20a7ff018c0","page":"representation-identical-inputs","operation":"move","beforeSHA256":"afe075866b6c3b9e472f28f019d9a27147050df06ab44305f6b2bde15e49f52a","afterSHA256":"afe075866b6c3b9e472f28f019d9a27147050df06ab44305f6b2bde15e49f52a","reason":"Same source body in authored semantic page"},{"oldId":"b-89c2b29f06e49fa493143cfa","targetId":"b-89c2b29f06e49fa493143cfa","page":"representation-identical-inputs","operation":"move","beforeSHA256":"87c14781e43c0cdb3d464384bdb5d7485899dfbba137dd9546fcdf80d049aa13","afterSHA256":"87c14781e43c0cdb3d464384bdb5d7485899dfbba137dd9546fcdf80d049aa13","reason":"Same source body in authored semantic page"},{"oldId":"b-a8cea04c9c47c43ca9054609","targetId":"b-a8cea04c9c47c43ca9054609","page":"representation-identical-inputs","operation":"move","beforeSHA256":"84e547945ec3d51d8e4451698ce1d34a4d11811ded25fe66e514a7e5f475e46f","afterSHA256":"84e547945ec3d51d8e4451698ce1d34a4d11811ded25fe66e514a7e5f475e46f","reason":"Same source body in authored semantic page"},{"oldId":"b-3eb05b399b178ab5fd8cfa0b","targetId":"b-3eb05b399b178ab5fd8cfa0b","page":"representation-reading-diagnosis","operation":"move","beforeSHA256":"5f302b801240bbed114e210e77627b73f44a32151f366ad74ddfdcae7da68f5e","afterSHA256":"5f302b801240bbed114e210e77627b73f44a32151f366ad74ddfdcae7da68f5e","reason":"Same source body in authored semantic page"},{"oldId":"b-b01849935f3d838bb26895c7","targetId":"b-b01849935f3d838bb26895c7","page":"representation-reading-diagnosis","operation":"move","beforeSHA256":"34391c7353cd0dc7b5aabaaf1cf7152abb3c8a4a23569a9796d889c7b4be3b1f","afterSHA256":"34391c7353cd0dc7b5aabaaf1cf7152abb3c8a4a23569a9796d889c7b4be3b1f","reason":"Same source body in authored semantic page"},{"oldId":"b-73530f3606906d0c65703e06","targetId":"b-73530f3606906d0c65703e06","page":"representation-reading-diagnosis","operation":"move","beforeSHA256":"91da615abd9f11d79355bc8fe5654d1246f1d2ef6db22daa46bc247bda588c02","afterSHA256":"91da615abd9f11d79355bc8fe5654d1246f1d2ef6db22daa46bc247bda588c02","reason":"Same source body in authored semantic page"},{"oldId":"b-6675bb1ed0895364f94d6f89","targetId":"b-6675bb1ed0895364f94d6f89","page":"representation-reading-diagnosis","operation":"move","beforeSHA256":"b27d134d90acacfbb8892413111edb846c4f494c80826d510b8f1d49ce8ef265","afterSHA256":"b27d134d90acacfbb8892413111edb846c4f494c80826d510b8f1d49ce8ef265","reason":"Same source body in authored semantic page"},{"oldId":"b-4123edfa246970a844542563","targetId":"b-4123edfa246970a844542563","page":"representation-reading-diagnosis","operation":"move","beforeSHA256":"1b4cb1d1397deda2ab0267237a6a26278baa4ad4eb3b946fac42318a852403d5","afterSHA256":"1b4cb1d1397deda2ab0267237a6a26278baa4ad4eb3b946fac42318a852403d5","reason":"Same source body in authored semantic page"},{"oldId":"b-081846a8ba6591e586f9bb10","targetId":"b-081846a8ba6591e586f9bb10","page":"representation-reading-diagnosis","operation":"move","beforeSHA256":"22e51595d9a7cf03b3165a3d14e280446eba8bcb2d5f98f6235bb0f8017a3033","afterSHA256":"22e51595d9a7cf03b3165a3d14e280446eba8bcb2d5f98f6235bb0f8017a3033","reason":"Same source body in authored semantic page"},{"oldId":"b-80f3c03d66899688b00d1301","targetId":"b-80f3c03d66899688b00d1301","page":"representation-reading-diagnosis","operation":"move","beforeSHA256":"9c6214cc5b5845cbab52bb501fd71ad4ef2eee82703ed0a35e7c6f6d184ed5e1","afterSHA256":"9c6214cc5b5845cbab52bb501fd71ad4ef2eee82703ed0a35e7c6f6d184ed5e1","reason":"Same source body in authored semantic page"},{"oldId":"b-a5a986e79095cb36ba4152bd","targetId":"b-a5a986e79095cb36ba4152bd","page":"representation-reading-diagnosis","operation":"move","beforeSHA256":"889643e455f0b36432752ae388e97264df6a2d7c156a17c9b26333c62a8c2257","afterSHA256":"889643e455f0b36432752ae388e97264df6a2d7c156a17c9b26333c62a8c2257","reason":"Same source body in authored semantic page"},{"oldId":"b-ea26ab3fc205a954b4d9fbeb","targetId":"b-ea26ab3fc205a954b4d9fbeb","page":"representation-reading-diagnosis","operation":"move","beforeSHA256":"cf44fcc62782fb4f6bdc42c16c5187d57556a03cb7ce6065e7bd9c36d8e7d733","afterSHA256":"cf44fcc62782fb4f6bdc42c16c5187d57556a03cb7ce6065e7bd9c36d8e7d733","reason":"Same source body in authored semantic page"},{"oldId":"b-89af65a34004b642f48f48dc","targetId":"b-89af65a34004b642f48f48dc","page":"representation-reading-diagnosis","operation":"move","beforeSHA256":"b8b69250624fd8c15e6f5db8690409848aaec027fc3b89222c6665c3044c9586","afterSHA256":"b8b69250624fd8c15e6f5db8690409848aaec027fc3b89222c6665c3044c9586","reason":"Same source body in authored semantic page"},{"oldId":"representation-identical-inputs-title","targetId":"b-303212a24f1c53a2f4627fe7","operation":"retain-source-identity"},{"oldId":"representation-reading-diagnosis-title","targetId":"b-4d4f2517e66d2b4bc5d0bb4a","operation":"retain-source-identity"},{"oldId":"representation-reading-diagnosis-bridge","targetId":"b-d0567c583bfb00b504ecbc8e","operation":"retain-source-identity"}],"blocks":[{"id":"b-303212a24f1c53a2f4627fe7","type":"h2","scope":"representation-identical-inputs","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p44.html#b-303212a24f1c53a2f4627fe7","content":"## 固定相同表示，推导为什么无法区分左右","sha256":"6b1b916cb5091eb32168e8c8f2c71306689fcc3df3b44b76b771ef6bda451ca2","chapter":{"id":"II","title":"区域、像素与实例"},"unit":{"id":"spatial-information-preservation","title":"瓶颈是读错位置，还是差别根本没有保留？"},"page":{"id":"representation-identical-inputs","title":"固定相同表示，推导为什么无法区分左右","number":1},"sectionTitle":"固定相同表示，推导为什么无法区分左右","currentQuestion":"怎样把信息碰撞的构造例与区域对齐的正面证据放在同一判断中？","sourceSlides":["l3-r2-p44"],"knowledgeBlock":null,"formulas":[],"previousBlock":null,"nextBlock":"b-5352ab09b69fecb5bf8bc01a","prerequisites":[{"id":"b-915b81c2c7d0b70935bc5355","url":"#b-915b81c2c7d0b70935bc5355","role":"problem-context"}],"relatedBlocks":[{"id":"b-d095f89c111a019d699824fd","url":"#b-d095f89c111a019d699824fd","relation":"figure-and-explanation"}]},{"id":"b-5352ab09b69fecb5bf8bc01a","type":"h2","scope":"explanation","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p44.html#b-5352ab09b69fecb5bf8bc01a","content":"## 解释、算例与证据","sha256":"d71cf15e707c8a3674b4c90a599efee783cb84da5b8c8236a1d01c3add7b4958","chapter":{"id":"II","title":"区域、像素与实例"},"unit":{"id":"spatial-information-preservation","title":"瓶颈是读错位置，还是差别根本没有保留？"},"page":{"id":"representation-identical-inputs","title":"固定相同表示，推导为什么无法区分左右","number":1},"sectionTitle":"固定相同表示，推导为什么无法区分左右","currentQuestion":"怎样把信息碰撞的构造例与区域对齐的正面证据放在同一判断中？","sourceSlides":["l3-r2-p44"],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-303212a24f1c53a2f4627fe7","nextBlock":"b-de79c86b47d5f73e04c8d7aa","prerequisites":[{"id":"b-303212a24f1c53a2f4627fe7","url":"#b-303212a24f1c53a2f4627fe7","role":"section-title"},{"id":"b-915b81c2c7d0b70935bc5355","url":"#b-915b81c2c7d0b70935bc5355","role":"problem-context"}],"relatedBlocks":[{"id":"b-d095f89c111a019d699824fd","url":"#b-d095f89c111a019d699824fd","relation":"figure-and-explanation"}]},{"id":"b-de79c86b47d5f73e04c8d7aa","type":"h1","scope":"document","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p44.html#b-de79c86b47d5f73e04c8d7aa","content":"# P44 · 两种布局，如果已经变成同一个数","sha256":"f4ec8b93ef44d945f955f22f4af5769481fd54084b82e0f15b7101aa9488a466","chapter":{"id":"II","title":"区域、像素与实例"},"unit":{"id":"spatial-information-preservation","title":"瓶颈是读错位置，还是差别根本没有保留？"},"page":{"id":"representation-identical-inputs","title":"固定相同表示，推导为什么无法区分左右","number":1},"sectionTitle":"固定相同表示，推导为什么无法区分左右","currentQuestion":"怎样把信息碰撞的构造例与区域对齐的正面证据放在同一判断中？","sourceSlides":["l3-r2-p44"],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-5352ab09b69fecb5bf8bc01a","nextBlock":"b-915b81c2c7d0b70935bc5355","prerequisites":[{"id":"b-5352ab09b69fecb5bf8bc01a","url":"#b-5352ab09b69fecb5bf8bc01a","role":"section-title"},{"id":"b-915b81c2c7d0b70935bc5355","url":"#b-915b81c2c7d0b70935bc5355","role":"problem-context"}],"relatedBlocks":[{"id":"b-d095f89c111a019d699824fd","url":"#b-d095f89c111a019d699824fd","relation":"figure-and-explanation"}]},{"id":"b-915b81c2c7d0b70935bc5355","type":"p","scope":"document","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p44.html#b-915b81c2c7d0b70935bc5355","content":"本页结论来自题设明确规定的相同表示、相同框和不能回看原图。确定性后处理收到相同输入只能产生相同输出。将其与论文训练对照并排，才能区分“读取已有差别”和“创造已不存在的差别”。","sha256":"bb805a18dc186d875e39921d6b91f9d80a17f61043f941f55e961c2edc4fad3e","chapter":{"id":"II","title":"区域、像素与实例"},"unit":{"id":"spatial-information-preservation","title":"瓶颈是读错位置，还是差别根本没有保留？"},"page":{"id":"representation-identical-inputs","title":"固定相同表示，推导为什么无法区分左右","number":1},"sectionTitle":"固定相同表示，推导为什么无法区分左右","currentQuestion":"怎样把信息碰撞的构造例与区域对齐的正面证据放在同一判断中？","sourceSlides":["l3-r2-p44"],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-de79c86b47d5f73e04c8d7aa","nextBlock":"b-d095f89c111a019d699824fd","prerequisites":[{"id":"b-de79c86b47d5f73e04c8d7aa","url":"#b-de79c86b47d5f73e04c8d7aa","role":"section-title"}],"relatedBlocks":[{"id":"b-d095f89c111a019d699824fd","url":"#b-d095f89c111a019d699824fd","relation":"figure-and-explanation"}]},{"id":"b-d095f89c111a019d699824fd","type":"figure","scope":"document","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p44.html#b-d095f89c111a019d699824fd","content":"图内标签与数值：Lecture 03 课堂图解 P44；A：亮在左列；1；0；1；0；B：亮在右列；0；1；0；1；整体均值；(1+0+1+0)/4；整体均值；(0+1+0+1)/4；0.5；0.5；同一个 F；同一个区域；同一种读取；不能区分 A / B\n\nA=[[1,0],[1,0]] 与 B=[[0,1],[0,1]] 分别取整体均值，汇入相同 0.5 后由同一后端处理。\n\n- A：亮在左列\n- 1\n- 0\n- 1\n- 0\n- B：亮在右列\n- 0\n- 1\n- 0\n- 1\n- 整体均值\n- (1+0+1+0)/4\n- 整体均值\n- (0+1+0+1)/4\n- 0.5\n- 0.5\n- 同一个 F\n- 同一个区域\n- 同一种读取\n- 不能区分 A / BA=[[1,0],[1,0]] 与 B=[[0,1],[0,1]] 分别取整体均值，汇入相同 0.5 后由同一后端处理。 [查看对应图解状态](/#/l3-r2-p44/1)[查看完整图解](/course/figures/l3-r2-p44-1.html)","sha256":"9a8dad14132213d371a29c72d35df52c915bb4e750b83704541f5e1846a4eed2","chapter":{"id":"II","title":"区域、像素与实例"},"unit":{"id":"spatial-information-preservation","title":"瓶颈是读错位置，还是差别根本没有保留？"},"page":{"id":"representation-identical-inputs","title":"固定相同表示，推导为什么无法区分左右","number":1},"sectionTitle":"固定相同表示，推导为什么无法区分左右","currentQuestion":"怎样把信息碰撞的构造例与区域对齐的正面证据放在同一判断中？","sourceSlides":["l3-r2-p44"],"knowledgeBlock":null,"formulas":[],"figure":{"caption":"A=[[1,0],[1,0]] 与 B=[[0,1],[0,1]] 分别取整体均值，汇入相同 0.5 后由同一后端处理。 查看对应图解状态","labels":["A：亮在左列","1","0","1","0","B：亮在右列","0","1","0","1","整体均值","(1+0+1+0)/4","整体均值","(0+1+0+1)/4","0.5","0.5","同一个 F","同一个区域","同一种读取","不能区分 A / B"],"description":"A=[[1,0],[1,0]] 与 B=[[0,1],[0,1]] 分别取整体均值，汇入相同 0.5 后由同一后端处理。A：亮在左列1010B：亮在右列0101整体均值(1+0+1+0)/4整体均值(0+1+0+1)/40.50.5同一个 F同一个区域同一种读取不能区分 A / B","evidenceIdentity":"A=[[1,0],[1,0]] 与 B=[[0,1],[0,1]] 分别取整体均值，汇入相同 0.5 后由同一后端处理。 查看对应图解状态","images":[]},"previousBlock":"b-915b81c2c7d0b70935bc5355","nextBlock":"b-09d657eb109005c80d10213b","prerequisites":[{"id":"b-de79c86b47d5f73e04c8d7aa","url":"#b-de79c86b47d5f73e04c8d7aa","role":"section-title"},{"id":"b-915b81c2c7d0b70935bc5355","url":"#b-915b81c2c7d0b70935bc5355","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-09d657eb109005c80d10213b","type":"h2","scope":"narrative","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p44.html#b-09d657eb109005c80d10213b","content":"## 课堂讲述","sha256":"786ce10a6b32a492259a3dccd0a54e61cd41d28a34382fb7b5386c5f03b7ad69","chapter":{"id":"II","title":"区域、像素与实例"},"unit":{"id":"spatial-information-preservation","title":"瓶颈是读错位置，还是差别根本没有保留？"},"page":{"id":"representation-identical-inputs","title":"固定相同表示，推导为什么无法区分左右","number":1},"sectionTitle":"固定相同表示，推导为什么无法区分左右","currentQuestion":"怎样把信息碰撞的构造例与区域对齐的正面证据放在同一判断中？","sourceSlides":["l3-r2-p44"],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-d095f89c111a019d699824fd","nextBlock":"b-b9ce927f3e29d58d7ef5e59a","prerequisites":[{"id":"b-de79c86b47d5f73e04c8d7aa","url":"#b-de79c86b47d5f73e04c8d7aa","role":"section-title"},{"id":"b-915b81c2c7d0b70935bc5355","url":"#b-915b81c2c7d0b70935bc5355","role":"problem-context"}],"relatedBlocks":[{"id":"b-d095f89c111a019d699824fd","url":"#b-d095f89c111a019d699824fd","relation":"figure-and-explanation"}]},{"id":"b-b9ce927f3e29d58d7ef5e59a","type":"p","scope":"narrative","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p44.html#b-b9ce927f3e29d58d7ef5e59a","content":"左边的亮像素都在左列，右边都在右列；如果任务要分辨左右，两种输入的答案不同。可是我们用整体均值作表示，两个结果都是零点五。\n现在冻结这个表示。下游拿到的 F 相同，区域也相同，只依赖这些信息的同一个确定性处理，就不可能知道原来是哪一种布局。多加几个输出格子，或者换一种插值，仍然是在处理同样的输入。\n这就是问卷第二题的核心。它不是在证明 RoIAlign 没用，而是在分清两个问题：读取位置错了，可以改读取；区别已经没保存，必须回到表示、输入或额外信息。相同输出不一定都来自同一种错误。","sha256":"7a149385c08b9a3671905a2798fb27bb47f83bba2b6d0b3940a4c3425c2117df","chapter":{"id":"II","title":"区域、像素与实例"},"unit":{"id":"spatial-information-preservation","title":"瓶颈是读错位置，还是差别根本没有保留？"},"page":{"id":"representation-identical-inputs","title":"固定相同表示，推导为什么无法区分左右","number":1},"sectionTitle":"固定相同表示，推导为什么无法区分左右","currentQuestion":"怎样把信息碰撞的构造例与区域对齐的正面证据放在同一判断中？","sourceSlides":["l3-r2-p44"],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-09d657eb109005c80d10213b","nextBlock":"b-7d77ceff9eb966c8abc9598e","prerequisites":[{"id":"b-09d657eb109005c80d10213b","url":"#b-09d657eb109005c80d10213b","role":"section-title"},{"id":"b-915b81c2c7d0b70935bc5355","url":"#b-915b81c2c7d0b70935bc5355","role":"problem-context"}],"relatedBlocks":[{"id":"b-d095f89c111a019d699824fd","url":"#b-d095f89c111a019d699824fd","relation":"figure-and-explanation"}]},{"id":"b-7d77ceff9eb966c8abc9598e","type":"h3","scope":"representation-collision-proof","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p44.html#b-7d77ceff9eb966c8abc9598e","content":"### 两个输入一旦碰撞，确定性读出为何无法分开","sha256":"b07830a9aa685e99757cc9b85c7a56ca2ea992afef9d9b9f8e88c83e45b6a619","chapter":{"id":"II","title":"区域、像素与实例"},"unit":{"id":"spatial-information-preservation","title":"瓶颈是读错位置，还是差别根本没有保留？"},"page":{"id":"representation-identical-inputs","title":"固定相同表示，推导为什么无法区分左右","number":1},"sectionTitle":"两个输入一旦碰撞，确定性读出为何无法分开","currentQuestion":"怎样把信息碰撞的构造例与区域对齐的正面证据放在同一判断中？","sourceSlides":["l3-r2-p44"],"knowledgeBlock":"representation-collision-proof","formulas":[],"previousBlock":"b-b9ce927f3e29d58d7ef5e59a","nextBlock":"b-31f7116af39dca2901da23fe","prerequisites":[{"id":"b-09d657eb109005c80d10213b","url":"#b-09d657eb109005c80d10213b","role":"section-title"},{"id":"b-915b81c2c7d0b70935bc5355","url":"#b-915b81c2c7d0b70935bc5355","role":"problem-context"}],"relatedBlocks":[{"id":"b-d095f89c111a019d699824fd","url":"#b-d095f89c111a019d699824fd","relation":"figure-and-explanation"},{"id":"b-31f7116af39dca2901da23fe","url":"#b-31f7116af39dca2901da23fe","relation":"same-reasoning"},{"id":"b-26caeac455cc1c2c6464ebf7","url":"#b-26caeac455cc1c2c6464ebf7","relation":"same-reasoning"},{"id":"b-a9a56bfbc46c888f0d322697","url":"#b-a9a56bfbc46c888f0d322697","relation":"same-reasoning"},{"id":"b-9cd63cc5f0f05ca3332d6784","url":"#b-9cd63cc5f0f05ca3332d6784","relation":"same-reasoning"}]},{"id":"b-31f7116af39dca2901da23fe","type":"p","scope":"representation-collision-proof","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p44.html#b-31f7116af39dca2901da23fe","content":"令甲 $A=[[1,0],[1,0]]$，乙 $B=[[0,1],[0,1]]$。如果编码器只取整体均值，$f(A)=f(B)=.5$。进一步固定框坐标、其他特征与读出参数，且读出不能再访问原图，那么对任何确定性函数 g，都有 $g(f(A),b)=g(f(B),b)$。","sha256":"bd243b07b2f5a86fb352c34ed9d40f637db6e093f7c59e8a243871e20a22699a","chapter":{"id":"II","title":"区域、像素与实例"},"unit":{"id":"spatial-information-preservation","title":"瓶颈是读错位置，还是差别根本没有保留？"},"page":{"id":"representation-identical-inputs","title":"固定相同表示，推导为什么无法区分左右","number":1},"sectionTitle":"两个输入一旦碰撞，确定性读出为何无法分开","currentQuestion":"怎样把信息碰撞的构造例与区域对齐的正面证据放在同一判断中？","sourceSlides":["l3-r2-p44"],"knowledgeBlock":"representation-collision-proof","formulas":[{"latex":"A=[[1,0],[1,0]]","definitionContext":"令甲 $A=[[1,0],[1,0]]$，乙 $B=[[0,1],[0,1]]$。如果编码器只取整体均值，$f(A)=f(B)=.5$。进一步固定框坐标、其他特征与读出参数，且读出不能再访问原图，那么对任何确定性函数 g，都有 $g(f(A),b)=g(f(B),b)$。\n这不是凭直觉说“分辨率低所以不行”，而是直接用相同输入必得相同输出。把 .5 复制为 4×4，或给同一个数做插值，都是对同样输入施加同样函数；结果仍无法告诉我们亮在左边还是右边。\n若允许随机性，同样条件下也只能得到相同的输出分布。偶尔猜对其中一个，并没有建立随输入布局变化而变化的判断。要系统地区分，就要重新提供差别：改变编码器保留的信息、引入另一层特征、允许回看原图，或给出不同的外部条件。\n题设没有证明真实粗特征普遍只等于均值，更没有推翻 RoIAlign 的完整模型结果。它是在迫使我们说清：性能瓶颈究竟是“已有信息读错位置”，还是“需要的信息根本没有保留”。"},{"latex":"B=[[0,1],[0,1]]","definitionContext":"令甲 $A=[[1,0],[1,0]]$，乙 $B=[[0,1],[0,1]]$。如果编码器只取整体均值，$f(A)=f(B)=.5$。进一步固定框坐标、其他特征与读出参数，且读出不能再访问原图，那么对任何确定性函数 g，都有 $g(f(A),b)=g(f(B),b)$。\n这不是凭直觉说“分辨率低所以不行”，而是直接用相同输入必得相同输出。把 .5 复制为 4×4，或给同一个数做插值，都是对同样输入施加同样函数；结果仍无法告诉我们亮在左边还是右边。\n若允许随机性，同样条件下也只能得到相同的输出分布。偶尔猜对其中一个，并没有建立随输入布局变化而变化的判断。要系统地区分，就要重新提供差别：改变编码器保留的信息、引入另一层特征、允许回看原图，或给出不同的外部条件。\n题设没有证明真实粗特征普遍只等于均值，更没有推翻 RoIAlign 的完整模型结果。它是在迫使我们说清：性能瓶颈究竟是“已有信息读错位置”，还是“需要的信息根本没有保留”。"},{"latex":"f(A)=f(B)=.5","definitionContext":"令甲 $A=[[1,0],[1,0]]$，乙 $B=[[0,1],[0,1]]$。如果编码器只取整体均值，$f(A)=f(B)=.5$。进一步固定框坐标、其他特征与读出参数，且读出不能再访问原图，那么对任何确定性函数 g，都有 $g(f(A),b)=g(f(B),b)$。\n这不是凭直觉说“分辨率低所以不行”，而是直接用相同输入必得相同输出。把 .5 复制为 4×4，或给同一个数做插值，都是对同样输入施加同样函数；结果仍无法告诉我们亮在左边还是右边。\n若允许随机性，同样条件下也只能得到相同的输出分布。偶尔猜对其中一个，并没有建立随输入布局变化而变化的判断。要系统地区分，就要重新提供差别：改变编码器保留的信息、引入另一层特征、允许回看原图，或给出不同的外部条件。\n题设没有证明真实粗特征普遍只等于均值，更没有推翻 RoIAlign 的完整模型结果。它是在迫使我们说清：性能瓶颈究竟是“已有信息读错位置”，还是“需要的信息根本没有保留”。"},{"latex":"g(f(A),b)=g(f(B),b)","definitionContext":"令甲 $A=[[1,0],[1,0]]$，乙 $B=[[0,1],[0,1]]$。如果编码器只取整体均值，$f(A)=f(B)=.5$。进一步固定框坐标、其他特征与读出参数，且读出不能再访问原图，那么对任何确定性函数 g，都有 $g(f(A),b)=g(f(B),b)$。\n这不是凭直觉说“分辨率低所以不行”，而是直接用相同输入必得相同输出。把 .5 复制为 4×4，或给同一个数做插值，都是对同样输入施加同样函数；结果仍无法告诉我们亮在左边还是右边。\n若允许随机性，同样条件下也只能得到相同的输出分布。偶尔猜对其中一个，并没有建立随输入布局变化而变化的判断。要系统地区分，就要重新提供差别：改变编码器保留的信息、引入另一层特征、允许回看原图，或给出不同的外部条件。\n题设没有证明真实粗特征普遍只等于均值，更没有推翻 RoIAlign 的完整模型结果。它是在迫使我们说清：性能瓶颈究竟是“已有信息读错位置”，还是“需要的信息根本没有保留”。"}],"previousBlock":"b-7d77ceff9eb966c8abc9598e","nextBlock":"b-26caeac455cc1c2c6464ebf7","prerequisites":[{"id":"b-7d77ceff9eb966c8abc9598e","url":"#b-7d77ceff9eb966c8abc9598e","role":"section-title"},{"id":"b-915b81c2c7d0b70935bc5355","url":"#b-915b81c2c7d0b70935bc5355","role":"problem-context"}],"relatedBlocks":[{"id":"b-d095f89c111a019d699824fd","url":"#b-d095f89c111a019d699824fd","relation":"figure-and-explanation"},{"id":"b-7d77ceff9eb966c8abc9598e","url":"#b-7d77ceff9eb966c8abc9598e","relation":"same-reasoning"},{"id":"b-26caeac455cc1c2c6464ebf7","url":"#b-26caeac455cc1c2c6464ebf7","relation":"same-reasoning"},{"id":"b-a9a56bfbc46c888f0d322697","url":"#b-a9a56bfbc46c888f0d322697","relation":"same-reasoning"},{"id":"b-9cd63cc5f0f05ca3332d6784","url":"#b-9cd63cc5f0f05ca3332d6784","relation":"same-reasoning"}]},{"id":"b-26caeac455cc1c2c6464ebf7","type":"p","scope":"representation-collision-proof","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p44.html#b-26caeac455cc1c2c6464ebf7","content":"这不是凭直觉说“分辨率低所以不行”，而是直接用相同输入必得相同输出。把 .5 复制为 4×4，或给同一个数做插值，都是对同样输入施加同样函数；结果仍无法告诉我们亮在左边还是右边。","sha256":"e77e8ad77aa56049b6fb8247fda69458bb1544e738be5d381781019ffb534dd2","chapter":{"id":"II","title":"区域、像素与实例"},"unit":{"id":"spatial-information-preservation","title":"瓶颈是读错位置，还是差别根本没有保留？"},"page":{"id":"representation-identical-inputs","title":"固定相同表示，推导为什么无法区分左右","number":1},"sectionTitle":"两个输入一旦碰撞，确定性读出为何无法分开","currentQuestion":"怎样把信息碰撞的构造例与区域对齐的正面证据放在同一判断中？","sourceSlides":["l3-r2-p44"],"knowledgeBlock":"representation-collision-proof","formulas":[],"previousBlock":"b-31f7116af39dca2901da23fe","nextBlock":"b-a9a56bfbc46c888f0d322697","prerequisites":[{"id":"b-7d77ceff9eb966c8abc9598e","url":"#b-7d77ceff9eb966c8abc9598e","role":"section-title"},{"id":"b-915b81c2c7d0b70935bc5355","url":"#b-915b81c2c7d0b70935bc5355","role":"problem-context"}],"relatedBlocks":[{"id":"b-d095f89c111a019d699824fd","url":"#b-d095f89c111a019d699824fd","relation":"figure-and-explanation"},{"id":"b-7d77ceff9eb966c8abc9598e","url":"#b-7d77ceff9eb966c8abc9598e","relation":"same-reasoning"},{"id":"b-31f7116af39dca2901da23fe","url":"#b-31f7116af39dca2901da23fe","relation":"same-reasoning"},{"id":"b-a9a56bfbc46c888f0d322697","url":"#b-a9a56bfbc46c888f0d322697","relation":"same-reasoning"},{"id":"b-9cd63cc5f0f05ca3332d6784","url":"#b-9cd63cc5f0f05ca3332d6784","relation":"same-reasoning"}]},{"id":"b-a9a56bfbc46c888f0d322697","type":"p","scope":"representation-collision-proof","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p44.html#b-a9a56bfbc46c888f0d322697","content":"若允许随机性，同样条件下也只能得到相同的输出分布。偶尔猜对其中一个，并没有建立随输入布局变化而变化的判断。要系统地区分，就要重新提供差别：改变编码器保留的信息、引入另一层特征、允许回看原图，或给出不同的外部条件。","sha256":"3d80b9ffcb056ac171153d7e1ef1eb3ad3c5e3d63645eab0b9ddd628ac06925a","chapter":{"id":"II","title":"区域、像素与实例"},"unit":{"id":"spatial-information-preservation","title":"瓶颈是读错位置，还是差别根本没有保留？"},"page":{"id":"representation-identical-inputs","title":"固定相同表示，推导为什么无法区分左右","number":1},"sectionTitle":"两个输入一旦碰撞，确定性读出为何无法分开","currentQuestion":"怎样把信息碰撞的构造例与区域对齐的正面证据放在同一判断中？","sourceSlides":["l3-r2-p44"],"knowledgeBlock":"representation-collision-proof","formulas":[],"previousBlock":"b-26caeac455cc1c2c6464ebf7","nextBlock":"b-9cd63cc5f0f05ca3332d6784","prerequisites":[{"id":"b-7d77ceff9eb966c8abc9598e","url":"#b-7d77ceff9eb966c8abc9598e","role":"section-title"},{"id":"b-915b81c2c7d0b70935bc5355","url":"#b-915b81c2c7d0b70935bc5355","role":"problem-context"}],"relatedBlocks":[{"id":"b-d095f89c111a019d699824fd","url":"#b-d095f89c111a019d699824fd","relation":"figure-and-explanation"},{"id":"b-7d77ceff9eb966c8abc9598e","url":"#b-7d77ceff9eb966c8abc9598e","relation":"same-reasoning"},{"id":"b-31f7116af39dca2901da23fe","url":"#b-31f7116af39dca2901da23fe","relation":"same-reasoning"},{"id":"b-26caeac455cc1c2c6464ebf7","url":"#b-26caeac455cc1c2c6464ebf7","relation":"same-reasoning"},{"id":"b-9cd63cc5f0f05ca3332d6784","url":"#b-9cd63cc5f0f05ca3332d6784","relation":"same-reasoning"}]},{"id":"b-9cd63cc5f0f05ca3332d6784","type":"p","scope":"representation-collision-proof","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p44.html#b-9cd63cc5f0f05ca3332d6784","content":"题设没有证明真实粗特征普遍只等于均值，更没有推翻 RoIAlign 的完整模型结果。它是在迫使我们说清：性能瓶颈究竟是“已有信息读错位置”，还是“需要的信息根本没有保留”。","sha256":"3de609c492ec57041f21576d7d8f5f74cfe093c085b75f1ea22b5d56980ed070","chapter":{"id":"II","title":"区域、像素与实例"},"unit":{"id":"spatial-information-preservation","title":"瓶颈是读错位置，还是差别根本没有保留？"},"page":{"id":"representation-identical-inputs","title":"固定相同表示，推导为什么无法区分左右","number":1},"sectionTitle":"两个输入一旦碰撞，确定性读出为何无法分开","currentQuestion":"怎样把信息碰撞的构造例与区域对齐的正面证据放在同一判断中？","sourceSlides":["l3-r2-p44"],"knowledgeBlock":"representation-collision-proof","formulas":[],"previousBlock":"b-a9a56bfbc46c888f0d322697","nextBlock":"b-f4217d5ed8c311ee2d519cbf","prerequisites":[{"id":"b-7d77ceff9eb966c8abc9598e","url":"#b-7d77ceff9eb966c8abc9598e","role":"section-title"},{"id":"b-915b81c2c7d0b70935bc5355","url":"#b-915b81c2c7d0b70935bc5355","role":"problem-context"}],"relatedBlocks":[{"id":"b-d095f89c111a019d699824fd","url":"#b-d095f89c111a019d699824fd","relation":"figure-and-explanation"},{"id":"b-7d77ceff9eb966c8abc9598e","url":"#b-7d77ceff9eb966c8abc9598e","relation":"same-reasoning"},{"id":"b-31f7116af39dca2901da23fe","url":"#b-31f7116af39dca2901da23fe","relation":"same-reasoning"},{"id":"b-26caeac455cc1c2c6464ebf7","url":"#b-26caeac455cc1c2c6464ebf7","relation":"same-reasoning"},{"id":"b-a9a56bfbc46c888f0d322697","url":"#b-a9a56bfbc46c888f0d322697","relation":"same-reasoning"}]},{"id":"b-f4217d5ed8c311ee2d519cbf","type":"h4","scope":"representation-identical-inputs","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p44.html#b-f4217d5ed8c311ee2d519cbf","content":"#### 图示条件与公式约定","sha256":"ef7f915c1ae63ed67bd23fecf050249273ba0efa8965d2c8d1c1c6b2d3178122","chapter":{"id":"II","title":"区域、像素与实例"},"unit":{"id":"spatial-information-preservation","title":"瓶颈是读错位置，还是差别根本没有保留？"},"page":{"id":"representation-identical-inputs","title":"固定相同表示，推导为什么无法区分左右","number":1},"sectionTitle":"固定相同表示，推导为什么无法区分左右","currentQuestion":"怎样把信息碰撞的构造例与区域对齐的正面证据放在同一判断中？","sourceSlides":["l3-r2-p44"],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-9cd63cc5f0f05ca3332d6784","nextBlock":"b-5a237dc21ad2fda4ff2cf1db","prerequisites":[{"id":"b-7d77ceff9eb966c8abc9598e","url":"#b-7d77ceff9eb966c8abc9598e","role":"section-title"},{"id":"b-915b81c2c7d0b70935bc5355","url":"#b-915b81c2c7d0b70935bc5355","role":"problem-context"}],"relatedBlocks":[{"id":"b-d095f89c111a019d699824fd","url":"#b-d095f89c111a019d699824fd","relation":"figure-and-explanation"}]},{"id":"b-5a237dc21ad2fda4ff2cf1db","type":"p","scope":"document","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p44.html#b-5a237dc21ad2fda4ff2cf1db","content":"构造的不可辨识例：固定确定性后端 g(F,RoI)，F 与 RoI 均相同，则 g 输出相同。随机后端的条件输出分布也相同。不能从此推出 RoIAlign 没有价值。正式 Tally 不变。","sha256":"59c8e843f3c5c93621033a4af327bb27e577a998f65f8b32398837089055ea5c","chapter":{"id":"II","title":"区域、像素与实例"},"unit":{"id":"spatial-information-preservation","title":"瓶颈是读错位置，还是差别根本没有保留？"},"page":{"id":"representation-identical-inputs","title":"固定相同表示，推导为什么无法区分左右","number":1},"sectionTitle":"固定相同表示，推导为什么无法区分左右","currentQuestion":"怎样把信息碰撞的构造例与区域对齐的正面证据放在同一判断中？","sourceSlides":["l3-r2-p44"],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-f4217d5ed8c311ee2d519cbf","nextBlock":"b-f7f633a33d4ecaff577710d3","prerequisites":[{"id":"b-f4217d5ed8c311ee2d519cbf","url":"#b-f4217d5ed8c311ee2d519cbf","role":"section-title"},{"id":"b-915b81c2c7d0b70935bc5355","url":"#b-915b81c2c7d0b70935bc5355","role":"problem-context"}],"relatedBlocks":[{"id":"b-d095f89c111a019d699824fd","url":"#b-d095f89c111a019d699824fd","relation":"figure-and-explanation"}]},{"id":"b-f7f633a33d4ecaff577710d3","type":"h4","scope":"retained-reading","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p44.html#b-f7f633a33d4ecaff577710d3","content":"#### 这条线索的展开与后续阅读","sha256":"48b8229f7559b36ed4d5628416e3991e5c73293fbce43374db67386c8736f3bc","chapter":{"id":"II","title":"区域、像素与实例"},"unit":{"id":"spatial-information-preservation","title":"瓶颈是读错位置，还是差别根本没有保留？"},"page":{"id":"representation-identical-inputs","title":"固定相同表示，推导为什么无法区分左右","number":1},"sectionTitle":"固定相同表示，推导为什么无法区分左右","currentQuestion":"怎样把信息碰撞的构造例与区域对齐的正面证据放在同一判断中？","sourceSlides":["l3-r2-p44"],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-5a237dc21ad2fda4ff2cf1db","nextBlock":"b-0cc7a89d266bb20a7ff018c0","prerequisites":[{"id":"b-f4217d5ed8c311ee2d519cbf","url":"#b-f4217d5ed8c311ee2d519cbf","role":"section-title"},{"id":"b-915b81c2c7d0b70935bc5355","url":"#b-915b81c2c7d0b70935bc5355","role":"problem-context"}],"relatedBlocks":[{"id":"b-d095f89c111a019d699824fd","url":"#b-d095f89c111a019d699824fd","relation":"figure-and-explanation"}]},{"id":"b-0cc7a89d266bb20a7ff018c0","type":"ul","scope":"retained-reading","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p44.html#b-0cc7a89d266bb20a7ff018c0","content":"- [左右不同，为什么都变成 0.5？](/supplement.html?return=l3-r2-p44/0#/l3-information-loss/0) 本页说明\n- [RoIAlign 修复读取错位，不修复表示丢失｜什么结论还不能推出](/supplement.html?return=l3-r2-p44/0#/l3-roi-align-chain-boundary/0) 本页说明","sha256":"afe075866b6c3b9e472f28f019d9a27147050df06ab44305f6b2bde15e49f52a","chapter":{"id":"II","title":"区域、像素与实例"},"unit":{"id":"spatial-information-preservation","title":"瓶颈是读错位置，还是差别根本没有保留？"},"page":{"id":"representation-identical-inputs","title":"固定相同表示，推导为什么无法区分左右","number":1},"sectionTitle":"固定相同表示，推导为什么无法区分左右","currentQuestion":"怎样把信息碰撞的构造例与区域对齐的正面证据放在同一判断中？","sourceSlides":["l3-r2-p44"],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-f7f633a33d4ecaff577710d3","nextBlock":"b-89c2b29f06e49fa493143cfa","prerequisites":[{"id":"b-f7f633a33d4ecaff577710d3","url":"#b-f7f633a33d4ecaff577710d3","role":"section-title"},{"id":"b-915b81c2c7d0b70935bc5355","url":"#b-915b81c2c7d0b70935bc5355","role":"problem-context"}],"relatedBlocks":[{"id":"b-d095f89c111a019d699824fd","url":"#b-d095f89c111a019d699824fd","relation":"figure-and-explanation"}]},{"id":"b-89c2b29f06e49fa493143cfa","type":"li","scope":"retained-reading","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p44.html#b-89c2b29f06e49fa493143cfa","content":"- [左右不同，为什么都变成 0.5？](/supplement.html?return=l3-r2-p44/0#/l3-information-loss/0) 本页说明","sha256":"87c14781e43c0cdb3d464384bdb5d7485899dfbba137dd9546fcdf80d049aa13","chapter":{"id":"II","title":"区域、像素与实例"},"unit":{"id":"spatial-information-preservation","title":"瓶颈是读错位置，还是差别根本没有保留？"},"page":{"id":"representation-identical-inputs","title":"固定相同表示，推导为什么无法区分左右","number":1},"sectionTitle":"固定相同表示，推导为什么无法区分左右","currentQuestion":"怎样把信息碰撞的构造例与区域对齐的正面证据放在同一判断中？","sourceSlides":["l3-r2-p44"],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-0cc7a89d266bb20a7ff018c0","nextBlock":"b-a8cea04c9c47c43ca9054609","prerequisites":[{"id":"b-f7f633a33d4ecaff577710d3","url":"#b-f7f633a33d4ecaff577710d3","role":"section-title"},{"id":"b-915b81c2c7d0b70935bc5355","url":"#b-915b81c2c7d0b70935bc5355","role":"problem-context"}],"relatedBlocks":[{"id":"b-d095f89c111a019d699824fd","url":"#b-d095f89c111a019d699824fd","relation":"figure-and-explanation"}]},{"id":"b-a8cea04c9c47c43ca9054609","type":"li","scope":"retained-reading","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p44.html#b-a8cea04c9c47c43ca9054609","content":"- [RoIAlign 修复读取错位，不修复表示丢失｜什么结论还不能推出](/supplement.html?return=l3-r2-p44/0#/l3-roi-align-chain-boundary/0) 本页说明","sha256":"84e547945ec3d51d8e4451698ce1d34a4d11811ded25fe66e514a7e5f475e46f","chapter":{"id":"II","title":"区域、像素与实例"},"unit":{"id":"spatial-information-preservation","title":"瓶颈是读错位置，还是差别根本没有保留？"},"page":{"id":"representation-identical-inputs","title":"固定相同表示，推导为什么无法区分左右","number":1},"sectionTitle":"固定相同表示，推导为什么无法区分左右","currentQuestion":"怎样把信息碰撞的构造例与区域对齐的正面证据放在同一判断中？","sourceSlides":["l3-r2-p44"],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-89c2b29f06e49fa493143cfa","nextBlock":null,"prerequisites":[{"id":"b-f7f633a33d4ecaff577710d3","url":"#b-f7f633a33d4ecaff577710d3","role":"section-title"},{"id":"b-915b81c2c7d0b70935bc5355","url":"#b-915b81c2c7d0b70935bc5355","role":"problem-context"}],"relatedBlocks":[{"id":"b-d095f89c111a019d699824fd","url":"#b-d095f89c111a019d699824fd","relation":"figure-and-explanation"}]},{"id":"b-4d4f2517e66d2b4bc5d0bb4a","type":"h2","scope":"representation-reading-diagnosis","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p44.html#b-4d4f2517e66d2b4bc5d0bb4a","content":"## 把对齐收益与信息限制放回各自条件","sha256":"eb532b3c5d4f75a628a2097bd4c044eb9b61f0a717ab4c3fae9f4a2659ea9cb8","chapter":{"id":"II","title":"区域、像素与实例"},"unit":{"id":"spatial-information-preservation","title":"瓶颈是读错位置，还是差别根本没有保留？"},"page":{"id":"representation-reading-diagnosis","title":"把对齐收益与信息限制放回各自条件","number":2},"sectionTitle":"把对齐收益与信息限制放回各自条件","currentQuestion":"怎样把信息碰撞的构造例与区域对齐的正面证据放在同一判断中？","sourceSlides":["l3-r2-p43","l3-r2-p44","l3-r2-p45"],"knowledgeBlock":null,"formulas":[],"previousBlock":null,"nextBlock":"b-d0567c583bfb00b504ecbc8e","prerequisites":[{"id":"b-d0567c583bfb00b504ecbc8e","url":"#b-d0567c583bfb00b504ecbc8e","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-d0567c583bfb00b504ecbc8e","type":"p","scope":"representation-reading-diagnosis","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p44.html#b-d0567c583bfb00b504ecbc8e","content":"这段模拟追问帮助核对两类实验的固定条件，不作为已发生课堂对话。","sha256":"b7f189442e3d137d26b346d7ec813cc0669a6c6938636af2ce91b91164568df1","chapter":{"id":"II","title":"区域、像素与实例"},"unit":{"id":"spatial-information-preservation","title":"瓶颈是读错位置，还是差别根本没有保留？"},"page":{"id":"representation-reading-diagnosis","title":"把对齐收益与信息限制放回各自条件","number":2},"sectionTitle":"把对齐收益与信息限制放回各自条件","currentQuestion":"怎样把信息碰撞的构造例与区域对齐的正面证据放在同一判断中？","sourceSlides":["l3-r2-p43","l3-r2-p44","l3-r2-p45"],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-4d4f2517e66d2b4bc5d0bb4a","nextBlock":"b-3eb05b399b178ab5fd8cfa0b","prerequisites":[{"id":"b-4d4f2517e66d2b4bc5d0bb4a","url":"#b-4d4f2517e66d2b4bc5d0bb4a","role":"section-title"}],"relatedBlocks":[]},{"id":"b-3eb05b399b178ab5fd8cfa0b","type":"h3","scope":"roi-information-dialogue","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p44.html#b-3eb05b399b178ab5fd8cfa0b","content":"### 备课推演：对齐有用，为什么问卷里又不能恢复？","sha256":"5f302b801240bbed114e210e77627b73f44a32151f366ad74ddfdcae7da68f5e","chapter":{"id":"II","title":"区域、像素与实例"},"unit":{"id":"spatial-information-preservation","title":"瓶颈是读错位置，还是差别根本没有保留？"},"page":{"id":"representation-reading-diagnosis","title":"把对齐收益与信息限制放回各自条件","number":2},"sectionTitle":"备课推演：对齐有用，为什么问卷里又不能恢复？","currentQuestion":"怎样把信息碰撞的构造例与区域对齐的正面证据放在同一判断中？","sourceSlides":["l3-r2-p43","l3-r2-p44","l3-r2-p45"],"knowledgeBlock":"roi-information-dialogue","formulas":[],"previousBlock":"b-d0567c583bfb00b504ecbc8e","nextBlock":"b-b01849935f3d838bb26895c7","prerequisites":[{"id":"b-4d4f2517e66d2b4bc5d0bb4a","url":"#b-4d4f2517e66d2b4bc5d0bb4a","role":"section-title"},{"id":"b-d0567c583bfb00b504ecbc8e","url":"#b-d0567c583bfb00b504ecbc8e","role":"problem-context"}],"relatedBlocks":[{"id":"b-b01849935f3d838bb26895c7","url":"#b-b01849935f3d838bb26895c7","relation":"same-reasoning"},{"id":"b-73530f3606906d0c65703e06","url":"#b-73530f3606906d0c65703e06","relation":"same-reasoning"},{"id":"b-6675bb1ed0895364f94d6f89","url":"#b-6675bb1ed0895364f94d6f89","relation":"same-reasoning"},{"id":"b-4123edfa246970a844542563","url":"#b-4123edfa246970a844542563","relation":"same-reasoning"},{"id":"b-081846a8ba6591e586f9bb10","url":"#b-081846a8ba6591e586f9bb10","relation":"same-reasoning"},{"id":"b-80f3c03d66899688b00d1301","url":"#b-80f3c03d66899688b00d1301","relation":"same-reasoning"},{"id":"b-a5a986e79095cb36ba4152bd","url":"#b-a5a986e79095cb36ba4152bd","relation":"same-reasoning"}]},{"id":"b-b01849935f3d838bb26895c7","type":"p","scope":"roi-information-dialogue","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p44.html#b-b01849935f3d838bb26895c7","content":"以下是备课用的模拟追问，不是已发生的课堂对话或学习效果记录。","sha256":"34391c7353cd0dc7b5aabaaf1cf7152abb3c8a4a23569a9796d889c7b4be3b1f","chapter":{"id":"II","title":"区域、像素与实例"},"unit":{"id":"spatial-information-preservation","title":"瓶颈是读错位置，还是差别根本没有保留？"},"page":{"id":"representation-reading-diagnosis","title":"把对齐收益与信息限制放回各自条件","number":2},"sectionTitle":"备课推演：对齐有用，为什么问卷里又不能恢复？","currentQuestion":"怎样把信息碰撞的构造例与区域对齐的正面证据放在同一判断中？","sourceSlides":["l3-r2-p43","l3-r2-p44","l3-r2-p45"],"knowledgeBlock":"roi-information-dialogue","formulas":[],"previousBlock":"b-3eb05b399b178ab5fd8cfa0b","nextBlock":"b-73530f3606906d0c65703e06","prerequisites":[{"id":"b-3eb05b399b178ab5fd8cfa0b","url":"#b-3eb05b399b178ab5fd8cfa0b","role":"section-title"},{"id":"b-d0567c583bfb00b504ecbc8e","url":"#b-d0567c583bfb00b504ecbc8e","role":"problem-context"}],"relatedBlocks":[{"id":"b-3eb05b399b178ab5fd8cfa0b","url":"#b-3eb05b399b178ab5fd8cfa0b","relation":"same-reasoning"},{"id":"b-73530f3606906d0c65703e06","url":"#b-73530f3606906d0c65703e06","relation":"same-reasoning"},{"id":"b-6675bb1ed0895364f94d6f89","url":"#b-6675bb1ed0895364f94d6f89","relation":"same-reasoning"},{"id":"b-4123edfa246970a844542563","url":"#b-4123edfa246970a844542563","relation":"same-reasoning"},{"id":"b-081846a8ba6591e586f9bb10","url":"#b-081846a8ba6591e586f9bb10","relation":"same-reasoning"},{"id":"b-80f3c03d66899688b00d1301","url":"#b-80f3c03d66899688b00d1301","relation":"same-reasoning"},{"id":"b-a5a986e79095cb36ba4152bd","url":"#b-a5a986e79095cb36ba4152bd","relation":"same-reasoning"}]},{"id":"b-73530f3606906d0c65703e06","type":"p","scope":"roi-information-dialogue","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p44.html#b-73530f3606906d0c65703e06","content":"**学生：** 论文说 RoIAlign 更准确，这里却说换采样没用，是不是自相矛盾？","sha256":"91da615abd9f11d79355bc8fe5654d1246f1d2ef6db22daa46bc247bda588c02","chapter":{"id":"II","title":"区域、像素与实例"},"unit":{"id":"spatial-information-preservation","title":"瓶颈是读错位置，还是差别根本没有保留？"},"page":{"id":"representation-reading-diagnosis","title":"把对齐收益与信息限制放回各自条件","number":2},"sectionTitle":"备课推演：对齐有用，为什么问卷里又不能恢复？","currentQuestion":"怎样把信息碰撞的构造例与区域对齐的正面证据放在同一判断中？","sourceSlides":["l3-r2-p43","l3-r2-p44","l3-r2-p45"],"knowledgeBlock":"roi-information-dialogue","formulas":[],"previousBlock":"b-b01849935f3d838bb26895c7","nextBlock":"b-6675bb1ed0895364f94d6f89","prerequisites":[{"id":"b-3eb05b399b178ab5fd8cfa0b","url":"#b-3eb05b399b178ab5fd8cfa0b","role":"section-title"},{"id":"b-d0567c583bfb00b504ecbc8e","url":"#b-d0567c583bfb00b504ecbc8e","role":"problem-context"}],"relatedBlocks":[{"id":"b-3eb05b399b178ab5fd8cfa0b","url":"#b-3eb05b399b178ab5fd8cfa0b","relation":"same-reasoning"},{"id":"b-b01849935f3d838bb26895c7","url":"#b-b01849935f3d838bb26895c7","relation":"same-reasoning"},{"id":"b-6675bb1ed0895364f94d6f89","url":"#b-6675bb1ed0895364f94d6f89","relation":"same-reasoning"},{"id":"b-4123edfa246970a844542563","url":"#b-4123edfa246970a844542563","relation":"same-reasoning"},{"id":"b-081846a8ba6591e586f9bb10","url":"#b-081846a8ba6591e586f9bb10","relation":"same-reasoning"},{"id":"b-80f3c03d66899688b00d1301","url":"#b-80f3c03d66899688b00d1301","relation":"same-reasoning"},{"id":"b-a5a986e79095cb36ba4152bd","url":"#b-a5a986e79095cb36ba4152bd","relation":"same-reasoning"}]},{"id":"b-6675bb1ed0895364f94d6f89","type":"p","scope":"roi-information-dialogue","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p44.html#b-6675bb1ed0895364f94d6f89","content":"**教师：** 先把两个输入条件指出来。论文比较不同区域读取方式训练的模型，特征里有可以利用的空间线索；正确坐标让它读得更好。现在图上甲乙都被我们规定为同一个 .5，框也相同，读取器没有得到任何左右差别。它不是读取错了地方，而是没有不同内容可读。","sha256":"b27d134d90acacfbb8892413111edb846c4f494c80826d510b8f1d49ce8ef265","chapter":{"id":"II","title":"区域、像素与实例"},"unit":{"id":"spatial-information-preservation","title":"瓶颈是读错位置，还是差别根本没有保留？"},"page":{"id":"representation-reading-diagnosis","title":"把对齐收益与信息限制放回各自条件","number":2},"sectionTitle":"备课推演：对齐有用，为什么问卷里又不能恢复？","currentQuestion":"怎样把信息碰撞的构造例与区域对齐的正面证据放在同一判断中？","sourceSlides":["l3-r2-p43","l3-r2-p44","l3-r2-p45"],"knowledgeBlock":"roi-information-dialogue","formulas":[],"previousBlock":"b-73530f3606906d0c65703e06","nextBlock":"b-4123edfa246970a844542563","prerequisites":[{"id":"b-3eb05b399b178ab5fd8cfa0b","url":"#b-3eb05b399b178ab5fd8cfa0b","role":"section-title"},{"id":"b-d0567c583bfb00b504ecbc8e","url":"#b-d0567c583bfb00b504ecbc8e","role":"problem-context"}],"relatedBlocks":[{"id":"b-3eb05b399b178ab5fd8cfa0b","url":"#b-3eb05b399b178ab5fd8cfa0b","relation":"same-reasoning"},{"id":"b-b01849935f3d838bb26895c7","url":"#b-b01849935f3d838bb26895c7","relation":"same-reasoning"},{"id":"b-73530f3606906d0c65703e06","url":"#b-73530f3606906d0c65703e06","relation":"same-reasoning"},{"id":"b-4123edfa246970a844542563","url":"#b-4123edfa246970a844542563","relation":"same-reasoning"},{"id":"b-081846a8ba6591e586f9bb10","url":"#b-081846a8ba6591e586f9bb10","relation":"same-reasoning"},{"id":"b-80f3c03d66899688b00d1301","url":"#b-80f3c03d66899688b00d1301","relation":"same-reasoning"},{"id":"b-a5a986e79095cb36ba4152bd","url":"#b-a5a986e79095cb36ba4152bd","relation":"same-reasoning"}]},{"id":"b-4123edfa246970a844542563","type":"p","scope":"roi-information-dialogue","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p44.html#b-4123edfa246970a844542563","content":"**学生：** 那我先把 .5 放大成很多像素呢？","sha256":"1b4cb1d1397deda2ab0267237a6a26278baa4ad4eb3b946fac42318a852403d5","chapter":{"id":"II","title":"区域、像素与实例"},"unit":{"id":"spatial-information-preservation","title":"瓶颈是读错位置，还是差别根本没有保留？"},"page":{"id":"representation-reading-diagnosis","title":"把对齐收益与信息限制放回各自条件","number":2},"sectionTitle":"备课推演：对齐有用，为什么问卷里又不能恢复？","currentQuestion":"怎样把信息碰撞的构造例与区域对齐的正面证据放在同一判断中？","sourceSlides":["l3-r2-p43","l3-r2-p44","l3-r2-p45"],"knowledgeBlock":"roi-information-dialogue","formulas":[],"previousBlock":"b-6675bb1ed0895364f94d6f89","nextBlock":"b-081846a8ba6591e586f9bb10","prerequisites":[{"id":"b-3eb05b399b178ab5fd8cfa0b","url":"#b-3eb05b399b178ab5fd8cfa0b","role":"section-title"},{"id":"b-d0567c583bfb00b504ecbc8e","url":"#b-d0567c583bfb00b504ecbc8e","role":"problem-context"}],"relatedBlocks":[{"id":"b-3eb05b399b178ab5fd8cfa0b","url":"#b-3eb05b399b178ab5fd8cfa0b","relation":"same-reasoning"},{"id":"b-b01849935f3d838bb26895c7","url":"#b-b01849935f3d838bb26895c7","relation":"same-reasoning"},{"id":"b-73530f3606906d0c65703e06","url":"#b-73530f3606906d0c65703e06","relation":"same-reasoning"},{"id":"b-6675bb1ed0895364f94d6f89","url":"#b-6675bb1ed0895364f94d6f89","relation":"same-reasoning"},{"id":"b-081846a8ba6591e586f9bb10","url":"#b-081846a8ba6591e586f9bb10","relation":"same-reasoning"},{"id":"b-80f3c03d66899688b00d1301","url":"#b-80f3c03d66899688b00d1301","relation":"same-reasoning"},{"id":"b-a5a986e79095cb36ba4152bd","url":"#b-a5a986e79095cb36ba4152bd","relation":"same-reasoning"}]},{"id":"b-081846a8ba6591e586f9bb10","type":"p","scope":"roi-information-dialogue","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p44.html#b-081846a8ba6591e586f9bb10","content":"**教师：** 指着甲乙放大的第一行看，它们是不是还一样？增加输出位置没有改变函数的输入。再看下一页，我们增加左右差通道，甲是 +1、乙是 −1。这次确实提供了新信息，所以同一个读出可以给出不同答案。","sha256":"22e51595d9a7cf03b3165a3d14e280446eba8bcb2d5f98f6235bb0f8017a3033","chapter":{"id":"II","title":"区域、像素与实例"},"unit":{"id":"spatial-information-preservation","title":"瓶颈是读错位置，还是差别根本没有保留？"},"page":{"id":"representation-reading-diagnosis","title":"把对齐收益与信息限制放回各自条件","number":2},"sectionTitle":"备课推演：对齐有用，为什么问卷里又不能恢复？","currentQuestion":"怎样把信息碰撞的构造例与区域对齐的正面证据放在同一判断中？","sourceSlides":["l3-r2-p43","l3-r2-p44","l3-r2-p45"],"knowledgeBlock":"roi-information-dialogue","formulas":[],"previousBlock":"b-4123edfa246970a844542563","nextBlock":"b-80f3c03d66899688b00d1301","prerequisites":[{"id":"b-3eb05b399b178ab5fd8cfa0b","url":"#b-3eb05b399b178ab5fd8cfa0b","role":"section-title"},{"id":"b-d0567c583bfb00b504ecbc8e","url":"#b-d0567c583bfb00b504ecbc8e","role":"problem-context"}],"relatedBlocks":[{"id":"b-3eb05b399b178ab5fd8cfa0b","url":"#b-3eb05b399b178ab5fd8cfa0b","relation":"same-reasoning"},{"id":"b-b01849935f3d838bb26895c7","url":"#b-b01849935f3d838bb26895c7","relation":"same-reasoning"},{"id":"b-73530f3606906d0c65703e06","url":"#b-73530f3606906d0c65703e06","relation":"same-reasoning"},{"id":"b-6675bb1ed0895364f94d6f89","url":"#b-6675bb1ed0895364f94d6f89","relation":"same-reasoning"},{"id":"b-4123edfa246970a844542563","url":"#b-4123edfa246970a844542563","relation":"same-reasoning"},{"id":"b-80f3c03d66899688b00d1301","url":"#b-80f3c03d66899688b00d1301","relation":"same-reasoning"},{"id":"b-a5a986e79095cb36ba4152bd","url":"#b-a5a986e79095cb36ba4152bd","relation":"same-reasoning"}]},{"id":"b-80f3c03d66899688b00d1301","type":"p","scope":"roi-information-dialogue","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p44.html#b-80f3c03d66899688b00d1301","content":"**学生：** 实际分割不好，我怎么知道是哪种情况？","sha256":"9c6214cc5b5845cbab52bb501fd71ad4ef2eee82703ed0a35e7c6f6d184ed5e1","chapter":{"id":"II","title":"区域、像素与实例"},"unit":{"id":"spatial-information-preservation","title":"瓶颈是读错位置，还是差别根本没有保留？"},"page":{"id":"representation-reading-diagnosis","title":"把对齐收益与信息限制放回各自条件","number":2},"sectionTitle":"备课推演：对齐有用，为什么问卷里又不能恢复？","currentQuestion":"怎样把信息碰撞的构造例与区域对齐的正面证据放在同一判断中？","sourceSlides":["l3-r2-p43","l3-r2-p44","l3-r2-p45"],"knowledgeBlock":"roi-information-dialogue","formulas":[],"previousBlock":"b-081846a8ba6591e586f9bb10","nextBlock":"b-a5a986e79095cb36ba4152bd","prerequisites":[{"id":"b-3eb05b399b178ab5fd8cfa0b","url":"#b-3eb05b399b178ab5fd8cfa0b","role":"section-title"},{"id":"b-d0567c583bfb00b504ecbc8e","url":"#b-d0567c583bfb00b504ecbc8e","role":"problem-context"}],"relatedBlocks":[{"id":"b-3eb05b399b178ab5fd8cfa0b","url":"#b-3eb05b399b178ab5fd8cfa0b","relation":"same-reasoning"},{"id":"b-b01849935f3d838bb26895c7","url":"#b-b01849935f3d838bb26895c7","relation":"same-reasoning"},{"id":"b-73530f3606906d0c65703e06","url":"#b-73530f3606906d0c65703e06","relation":"same-reasoning"},{"id":"b-6675bb1ed0895364f94d6f89","url":"#b-6675bb1ed0895364f94d6f89","relation":"same-reasoning"},{"id":"b-4123edfa246970a844542563","url":"#b-4123edfa246970a844542563","relation":"same-reasoning"},{"id":"b-081846a8ba6591e586f9bb10","url":"#b-081846a8ba6591e586f9bb10","relation":"same-reasoning"},{"id":"b-a5a986e79095cb36ba4152bd","url":"#b-a5a986e79095cb36ba4152bd","relation":"same-reasoning"}]},{"id":"b-a5a986e79095cb36ba4152bd","type":"p","scope":"roi-information-dialogue","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p44.html#b-a5a986e79095cb36ba4152bd","content":"**教师：** 先检查标注与坐标能否对上，再固定模型做采样对照；若疑似特征不区分，可以尝试接入较细层，或设计有针对性的 probe。一次失败可能有多处原因，不能从一张模糊 mask 直接断言。","sha256":"889643e455f0b36432752ae388e97264df6a2d7c156a17c9b26333c62a8c2257","chapter":{"id":"II","title":"区域、像素与实例"},"unit":{"id":"spatial-information-preservation","title":"瓶颈是读错位置，还是差别根本没有保留？"},"page":{"id":"representation-reading-diagnosis","title":"把对齐收益与信息限制放回各自条件","number":2},"sectionTitle":"备课推演：对齐有用，为什么问卷里又不能恢复？","currentQuestion":"怎样把信息碰撞的构造例与区域对齐的正面证据放在同一判断中？","sourceSlides":["l3-r2-p43","l3-r2-p44","l3-r2-p45"],"knowledgeBlock":"roi-information-dialogue","formulas":[],"previousBlock":"b-80f3c03d66899688b00d1301","nextBlock":"b-ea26ab3fc205a954b4d9fbeb","prerequisites":[{"id":"b-3eb05b399b178ab5fd8cfa0b","url":"#b-3eb05b399b178ab5fd8cfa0b","role":"section-title"},{"id":"b-d0567c583bfb00b504ecbc8e","url":"#b-d0567c583bfb00b504ecbc8e","role":"problem-context"}],"relatedBlocks":[{"id":"b-3eb05b399b178ab5fd8cfa0b","url":"#b-3eb05b399b178ab5fd8cfa0b","relation":"same-reasoning"},{"id":"b-b01849935f3d838bb26895c7","url":"#b-b01849935f3d838bb26895c7","relation":"same-reasoning"},{"id":"b-73530f3606906d0c65703e06","url":"#b-73530f3606906d0c65703e06","relation":"same-reasoning"},{"id":"b-6675bb1ed0895364f94d6f89","url":"#b-6675bb1ed0895364f94d6f89","relation":"same-reasoning"},{"id":"b-4123edfa246970a844542563","url":"#b-4123edfa246970a844542563","relation":"same-reasoning"},{"id":"b-081846a8ba6591e586f9bb10","url":"#b-081846a8ba6591e586f9bb10","relation":"same-reasoning"},{"id":"b-80f3c03d66899688b00d1301","url":"#b-80f3c03d66899688b00d1301","relation":"same-reasoning"}]},{"id":"b-ea26ab3fc205a954b4d9fbeb","type":"h2","scope":"practice","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p44.html#b-ea26ab3fc205a954b4d9fbeb","content":"## 沿这个问题继续","sha256":"cf44fcc62782fb4f6bdc42c16c5187d57556a03cb7ce6065e7bd9c36d8e7d733","chapter":{"id":"II","title":"区域、像素与实例"},"unit":{"id":"spatial-information-preservation","title":"瓶颈是读错位置，还是差别根本没有保留？"},"page":{"id":"representation-reading-diagnosis","title":"把对齐收益与信息限制放回各自条件","number":2},"sectionTitle":"把对齐收益与信息限制放回各自条件","currentQuestion":"怎样把信息碰撞的构造例与区域对齐的正面证据放在同一判断中？","sourceSlides":["l3-r2-p43","l3-r2-p44","l3-r2-p45"],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-a5a986e79095cb36ba4152bd","nextBlock":"b-89af65a34004b642f48f48dc","prerequisites":[{"id":"b-3eb05b399b178ab5fd8cfa0b","url":"#b-3eb05b399b178ab5fd8cfa0b","role":"section-title"},{"id":"b-d0567c583bfb00b504ecbc8e","url":"#b-d0567c583bfb00b504ecbc8e","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-89af65a34004b642f48f48dc","type":"p","scope":"practice","url":"https://codingai-lec03.pages.dev/course/pages/l3-r2-p44.html#b-89af65a34004b642f48f48dc","content":"先读本页解释与图，再到 LAB03 改一个坐标或对象。","sha256":"b8b69250624fd8c15e6f5db8690409848aaec027fc3b89222c6665c3044c9586","chapter":{"id":"II","title":"区域、像素与实例"},"unit":{"id":"spatial-information-preservation","title":"瓶颈是读错位置，还是差别根本没有保留？"},"page":{"id":"representation-reading-diagnosis","title":"把对齐收益与信息限制放回各自条件","number":2},"sectionTitle":"把对齐收益与信息限制放回各自条件","currentQuestion":"怎样把信息碰撞的构造例与区域对齐的正面证据放在同一判断中？","sourceSlides":["l3-r2-p43","l3-r2-p44","l3-r2-p45"],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-ea26ab3fc205a954b4d9fbeb","nextBlock":null,"prerequisites":[{"id":"b-ea26ab3fc205a954b4d9fbeb","url":"#b-ea26ab3fc205a954b4d9fbeb","role":"section-title"},{"id":"b-d0567c583bfb00b504ecbc8e","url":"#b-d0567c583bfb00b504ecbc8e","role":"problem-context"}],"relatedBlocks":[]}]}