http://www.pikafish.com/wiki/index.php?action=history&feed=atom&title=%E8%B1%A1%E6%A3%8B%E6%9C%89%E2%80%9C%E9%98%BF%E5%B0%94%E6%B3%95%E7%8B%97%E2%80%9D%E5%90%97%EF%BC%9F 2026-04-29T21:49:33Z 本wiki上该页面的版本历史 MediaWiki 1.40.0 http://www.pikafish.com/wiki/index.php?title=%E8%B1%A1%E6%A3%8B%E6%9C%89%E2%80%9C%E9%98%BF%E5%B0%94%E6%B3%95%E7%8B%97%E2%80%9D%E5%90%97%EF%BC%9F&diff=482&oldid=prev 2025-01-05T23:59:58Z

<p></p> <table style="background-color: #fff; color: #202122;" data-mw="interface"> <col class="diff-marker" /> <col class="diff-content" /> <col class="diff-marker" /> <col class="diff-content" /> <tr class="diff-title" lang="zh-Hans-CN"> <td colspan="2" style="background-color: #fff; color: #202122; text-align: center;">←上一版本</td> <td colspan="2" style="background-color: #fff; color: #202122; text-align: center;">2025年1月6日 (一) 07:59的版本</td> </tr><tr><td colspan="2" class="diff-lineno" id="mw-diff-left-l2">第2行：</td> <td colspan="2" class="diff-lineno">第2行：</td></tr> <tr><td class="diff-marker"></td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>&lt;br&gt;</div></td><td class="diff-marker"></td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>&lt;br&gt;</div></td></tr> <tr><td class="diff-marker"></td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>&lt;br&gt;</div></td><td class="diff-marker"></td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>&lt;br&gt;</div></td></tr> <tr><td class="diff-marker" data-marker="−"></td><td style="color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;"><div>没有，但有相同算法的px0，官网[http://px0.org px0.org]<del style="font-weight: bold; text-decoration: none;">（QQ群号578342679）</del></div></td><td class="diff-marker" data-marker="+"></td><td style="color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;"><div>没有，但有相同算法的px0，官网[http://px0.org px0.org]<ins style="font-weight: bold; text-decoration: none;">，但目前看在象棋领域，CPU引擎仍然强于显卡引擎。</ins></div></td></tr> <tr><td class="diff-marker"></td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><br/></td><td class="diff-marker"></td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><br/></td></tr> <tr><td class="diff-marker"></td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><br/></td><td class="diff-marker"></td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><br/></td></tr> <tr><td class="diff-marker" data-marker="−"></td><td style="color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;"><div>多年前的阿尔法狗、阿尔法零（AlphaGo AlphaZero)<del style="font-weight: bold; text-decoration: none;">用的是另外一种搜索算法，用来配合大神经网络架构。</del></div></td><td class="diff-marker" data-marker="+"></td><td style="color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;"><div>多年前的阿尔法狗、阿尔法零（AlphaGo AlphaZero)<ins style="font-weight: bold; text-decoration: none;">用的是另外一种搜索算法，用来配合真正的重点：卷积神经网络架构。</ins></div></td></tr> <tr><td class="diff-marker"></td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>这种引擎通常主要使用显卡(GPU)或者TPU。</div></td><td class="diff-marker"></td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>这种引擎通常主要使用显卡(GPU)或者TPU。</div></td></tr> <tr><td class="diff-marker"></td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>而象棋强引擎的搜索算法都只需要CPU，不需要显卡。</div></td><td class="diff-marker"></td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>而象棋强引擎的搜索算法都只需要CPU，不需要显卡。</div></td></tr> <tr><td colspan="2" class="diff-lineno" id="mw-diff-left-l17">第17行：</td> <td colspan="2" class="diff-lineno">第17行：</td></tr> <tr><td class="diff-marker"></td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>现在的顶尖围棋、国象、日将引擎棋力已经远超当年的阿尔法零。</div></td><td class="diff-marker"></td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>现在的顶尖围棋、国象、日将引擎棋力已经远超当年的阿尔法零。</div></td></tr> <tr><td class="diff-marker"></td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><br/></td><td class="diff-marker"></td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><br/></td></tr> <tr><td class="diff-marker" data-marker="−"></td><td style="color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;"><div><del style="font-weight: bold; text-decoration: none;">对显卡引擎来说，从随机落子开始迭代训练，其上限高于用其他引擎或者人类数据训练的。AlphaGoZero便是如此，其棋力远高于与李世石对弈的AlphaGoLee(从人类数据开始迭代训练)。</del></div></td><td class="diff-marker" data-marker="+"></td><td style="color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;"><div><ins style="font-weight: bold; text-decoration: none;">对显卡引擎来说，从随机落子开始迭代训练和从人类数据开始迭代训练，两者的上限是否有区别依然存疑，从随机落子为起点可能只是噱头，毕竟a0和ag0还有其他地方的差异。</ins></div></td></tr> <tr><td class="diff-marker"></td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><br/></td><td class="diff-marker"></td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><br/></td></tr> <tr><td class="diff-marker"></td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>需要注意的是，阿尔法猫使用的算法和AlphaGo不同，属于ab剪枝，仅仅是名字相似。</div></td><td class="diff-marker"></td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>需要注意的是，阿尔法猫使用的算法和AlphaGo不同，属于ab剪枝，仅仅是名字相似。</div></td></tr> <!-- diff cache key pikafish_wiki-pkf:diff::1.12:old-260:rev-482 --> </table>

Afkbad http://www.pikafish.com/wiki/index.php?title=%E8%B1%A1%E6%A3%8B%E6%9C%89%E2%80%9C%E9%98%BF%E5%B0%94%E6%B3%95%E7%8B%97%E2%80%9D%E5%90%97%EF%BC%9F&diff=260&oldid=prev 2024-03-31T15:03:16Z

<p></p> <table style="background-color: #fff; color: #202122;" data-mw="interface"> <col class="diff-marker" /> <col class="diff-content" /> <col class="diff-marker" /> <col class="diff-content" /> <tr class="diff-title" lang="zh-Hans-CN"> <td colspan="2" style="background-color: #fff; color: #202122; text-align: center;">←上一版本</td> <td colspan="2" style="background-color: #fff; color: #202122; text-align: center;">2024年3月31日 (日) 23:03的版本</td> </tr><tr><td colspan="2" class="diff-lineno" id="mw-diff-left-l1">第1行：</td> <td colspan="2" class="diff-lineno">第1行：</td></tr> <tr><td class="diff-marker" data-marker="−"></td><td style="color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;"><div><del style="font-weight: bold; text-decoration: none;">有相同算法的px0，官网</del>[http://px0.org px0.org]（QQ群号578342679）</div></td><td class="diff-marker" data-marker="+"></td><td style="color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;"><div><ins style="font-weight: bold; text-decoration: none;">&lt;small&gt;[[棋软知识|返回“棋软知识”]]&lt;/small&gt;</ins></div></td></tr> <tr><td colspan="2" class="diff-side-deleted"></td><td class="diff-marker" data-marker="+"></td><td style="color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;"><div><ins style="font-weight: bold; text-decoration: none;">&lt;br&gt;</ins></div></td></tr> <tr><td colspan="2" class="diff-side-deleted"></td><td class="diff-marker" data-marker="+"></td><td style="color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;"><div><ins style="font-weight: bold; text-decoration: none;">&lt;br&gt;</ins></div></td></tr> <tr><td colspan="2" class="diff-side-deleted"></td><td class="diff-marker" data-marker="+"></td><td style="color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;"><div><ins style="font-weight: bold; text-decoration: none;">没有，但有相同算法的px0，官网</ins>[http://px0.org px0.org]（QQ群号578342679）</div></td></tr> <tr><td class="diff-marker"></td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><br/></td><td class="diff-marker"></td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><br/></td></tr> <tr><td class="diff-marker"></td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><br/></td><td class="diff-marker"></td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><br/></td></tr> <!-- diff cache key pikafish_wiki-pkf:diff::1.12:old-118:rev-260 --> </table>

Afkbad http://www.pikafish.com/wiki/index.php?title=%E8%B1%A1%E6%A3%8B%E6%9C%89%E2%80%9C%E9%98%BF%E5%B0%94%E6%B3%95%E7%8B%97%E2%80%9D%E5%90%97%EF%BC%9F&diff=118&oldid=prev 2024-02-18T14:30:14Z

<p></p> <table style="background-color: #fff; color: #202122;" data-mw="interface"> <col class="diff-marker" /> <col class="diff-content" /> <col class="diff-marker" /> <col class="diff-content" /> <tr class="diff-title" lang="zh-Hans-CN"> <td colspan="2" style="background-color: #fff; color: #202122; text-align: center;">←上一版本</td> <td colspan="2" style="background-color: #fff; color: #202122; text-align: center;">2024年2月18日 (日) 22:30的版本</td> </tr><tr><td colspan="2" class="diff-lineno" id="mw-diff-left-l1">第1行：</td> <td colspan="2" class="diff-lineno">第1行：</td></tr> <tr><td class="diff-marker" data-marker="−"></td><td style="color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;"><div>有相同算法的px0，官网[http://px0.org px0.org]</div></td><td class="diff-marker" data-marker="+"></td><td style="color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;"><div>有相同算法的px0，官网[http://px0.org px0.org]<ins style="font-weight: bold; text-decoration: none;">（QQ群号578342679）</ins></div></td></tr> <tr><td class="diff-marker"></td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><br/></td><td class="diff-marker"></td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><br/></td></tr> <tr><td class="diff-marker"></td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><br/></td><td class="diff-marker"></td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><br/></td></tr> <tr><td colspan="2" class="diff-lineno" id="mw-diff-left-l15">第15行：</td> <td colspan="2" class="diff-lineno">第15行：</td></tr> <tr><td class="diff-marker"></td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><br/></td><td class="diff-marker"></td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><br/></td></tr> <tr><td class="diff-marker"></td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>对显卡引擎来说，从随机落子开始迭代训练，其上限高于用其他引擎或者人类数据训练的。AlphaGoZero便是如此，其棋力远高于与李世石对弈的AlphaGoLee(从人类数据开始迭代训练)。</div></td><td class="diff-marker"></td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>对显卡引擎来说，从随机落子开始迭代训练，其上限高于用其他引擎或者人类数据训练的。AlphaGoZero便是如此，其棋力远高于与李世石对弈的AlphaGoLee(从人类数据开始迭代训练)。</div></td></tr> <tr><td colspan="2" class="diff-side-deleted"></td><td class="diff-marker" data-marker="+"></td><td style="color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;"><div><ins style="font-weight: bold; text-decoration: none;"></ins></div></td></tr> <tr><td colspan="2" class="diff-side-deleted"></td><td class="diff-marker" data-marker="+"></td><td style="color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;"><div><ins style="font-weight: bold; text-decoration: none;">需要注意的是，阿尔法猫使用的算法和AlphaGo不同，属于ab剪枝，仅仅是名字相似。</ins></div></td></tr> <!-- diff cache key pikafish_wiki-pkf:diff::1.12:old-33:rev-118 --> </table>

New http://www.pikafish.com/wiki/index.php?title=%E8%B1%A1%E6%A3%8B%E6%9C%89%E2%80%9C%E9%98%BF%E5%B0%94%E6%B3%95%E7%8B%97%E2%80%9D%E5%90%97%EF%BC%9F&diff=33&oldid=prev 2023-11-22T19:13:25Z

<p></p> <table style="background-color: #fff; color: #202122;" data-mw="interface"> <col class="diff-marker" /> <col class="diff-content" /> <col class="diff-marker" /> <col class="diff-content" /> <tr class="diff-title" lang="zh-Hans-CN"> <td colspan="2" style="background-color: #fff; color: #202122; text-align: center;">←上一版本</td> <td colspan="2" style="background-color: #fff; color: #202122; text-align: center;">2023年11月23日 (四) 03:13的版本</td> </tr><tr><td colspan="2" class="diff-lineno" id="mw-diff-left-l1">第1行：</td> <td colspan="2" class="diff-lineno">第1行：</td></tr> <tr><td class="diff-marker"></td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>有相同算法的px0，官网[http://px0.org px0.org]</div></td><td class="diff-marker"></td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>有相同算法的px0，官网[http://px0.org px0.org]</div></td></tr> <tr><td colspan="2" class="diff-side-deleted"></td><td class="diff-marker" data-marker="+"></td><td style="color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;"><div><ins style="font-weight: bold; text-decoration: none;"></ins></div></td></tr> <tr><td class="diff-marker"></td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><br/></td><td class="diff-marker"></td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><br/></td></tr> <tr><td class="diff-marker"></td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>多年前的阿尔法狗、阿尔法零（AlphaGo AlphaZero)用的是另外一种搜索算法，用来配合大神经网络架构。</div></td><td class="diff-marker"></td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>多年前的阿尔法狗、阿尔法零（AlphaGo AlphaZero)用的是另外一种搜索算法，用来配合大神经网络架构。</div></td></tr> <!-- diff cache key pikafish_wiki-pkf:diff::1.12:old-32:rev-33 --> </table>

Afkbad http://www.pikafish.com/wiki/index.php?title=%E8%B1%A1%E6%A3%8B%E6%9C%89%E2%80%9C%E9%98%BF%E5%B0%94%E6%B3%95%E7%8B%97%E2%80%9D%E5%90%97%EF%BC%9F&diff=32&oldid=prev 2023-11-22T19:13:10Z

<p></p> <table style="background-color: #fff; color: #202122;" data-mw="interface"> <col class="diff-marker" /> <col class="diff-content" /> <col class="diff-marker" /> <col class="diff-content" /> <tr class="diff-title" lang="zh-Hans-CN"> <td colspan="2" style="background-color: #fff; color: #202122; text-align: center;">←上一版本</td> <td colspan="2" style="background-color: #fff; color: #202122; text-align: center;">2023年11月23日 (四) 03:13的版本</td> </tr><tr><td colspan="2" class="diff-lineno" id="mw-diff-left-l1">第1行：</td> <td colspan="2" class="diff-lineno">第1行：</td></tr> <tr><td colspan="2" class="diff-side-deleted"></td><td class="diff-marker" data-marker="+"></td><td style="color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;"><div><ins style="font-weight: bold; text-decoration: none;">有相同算法的px0，官网[http://px0.org px0.org]</ins></div></td></tr> <tr><td colspan="2" class="diff-side-deleted"></td><td class="diff-marker" data-marker="+"></td><td style="color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;"><div><ins style="font-weight: bold; text-decoration: none;"></ins></div></td></tr> <tr><td class="diff-marker"></td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>多年前的阿尔法狗、阿尔法零（AlphaGo AlphaZero)用的是另外一种搜索算法，用来配合大神经网络架构。</div></td><td class="diff-marker"></td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>多年前的阿尔法狗、阿尔法零（AlphaGo AlphaZero)用的是另外一种搜索算法，用来配合大神经网络架构。</div></td></tr> <tr><td class="diff-marker"></td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>这种引擎通常主要使用显卡(GPU)或者TPU。</div></td><td class="diff-marker"></td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>这种引擎通常主要使用显卡(GPU)或者TPU。</div></td></tr> <!-- diff cache key pikafish_wiki-pkf:diff::1.12:old-31:rev-32 --> </table>

Afkbad http://www.pikafish.com/wiki/index.php?title=%E8%B1%A1%E6%A3%8B%E6%9C%89%E2%80%9C%E9%98%BF%E5%B0%94%E6%B3%95%E7%8B%97%E2%80%9D%E5%90%97%EF%BC%9F&diff=31&oldid=prev 2023-11-22T19:11:53Z

<p>创建页面，内容为“多年前的阿尔法狗、阿尔法零（AlphaGo AlphaZero)用的是另外一种搜索算法，用来配合大神经网络架构。 这种引擎通常主要使用显卡(GPU)或者TPU。 而象棋强引擎的搜索算法都只需要CPU，不需要显卡。 显卡引擎的网络足够大，所以评估更加准确。和nnue不同，显卡引擎的网络不止评估网络，还包含了策略网络，策略网络负责决定搜索着法，类似于人类看到…”</p> <p><b>新页面</b></p><div>多年前的阿尔法狗、阿尔法零（AlphaGo AlphaZero)用的是另外一种搜索算法，用来配合大神经网络架构。<br /> 这种引擎通常主要使用显卡(GPU)或者TPU。<br /> 而象棋强引擎的搜索算法都只需要CPU，不需要显卡。<br /> <br /> <br /> 显卡引擎的网络足够大，所以评估更加准确。和nnue不同，显卡引擎的网络不止评估网络，还包含了策略网络，策略网络负责决定搜索着法，类似于人类看到一个局面，想要进行计算，就必须先选择一个着法进行计算，如果这种“直觉”够准，可以节省大量的时间。<br /> 但这种大架构让显卡引擎更依赖显卡和mcts搜索(实际上棋类显卡引擎已经不是狭义上的MC蒙特卡洛了)。<br /> 目前，显卡引擎在象棋和国际象棋上，仍未超越CPU引擎。<br /> 阿尔法系列并未开源，围棋的katago leela、国际象棋的lc0都是根据其论文复现的，而象棋有克隆lc0的px0、ggz。<br /> <br /> 现在的顶尖围棋、国象、日将引擎棋力已经远超当年的阿尔法零。<br /> <br /> 对显卡引擎来说，从随机落子开始迭代训练，其上限高于用其他引擎或者人类数据训练的。AlphaGoZero便是如此，其棋力远高于与李世石对弈的AlphaGoLee(从人类数据开始迭代训练)。</div>

Afkbad