{"id":477510,"date":"2023-08-09T09:15:57","date_gmt":"2023-08-09T09:15:57","guid":{"rendered":""},"modified":"2023-09-05T11:14:51","modified_gmt":"2023-09-05T11:14:51","slug":"human-in-the-loop","status":"publish","type":"wiki","link":"https:\/\/oneproxy.pro\/cn\/wiki\/human-in-the-loop\/","title":{"rendered":"\u4eba\u673a\u4ea4\u4e92"},"content":{"rendered":"<p>\u4eba\u5728\u73af (HITL) \u662f\u4e00\u79cd\u4ea4\u4e92\u5f0f\u8ba1\u7b97\u65b9\u6cd5\uff0c\u5b83\u5c06\u4eba\u7c7b\u667a\u80fd\u4e0e\u4eba\u5de5\u667a\u80fd (AI) \u7cfb\u7edf\u76f8\u96c6\u6210\uff0c\u4ee5\u66f4\u9ad8\u6548\u3001\u66f4\u51c6\u786e\u5730\u5b8c\u6210\u4efb\u52a1\u3002<\/p>\n<h2>\u4eba\u5728\u73af\u7684\u8d77\u6e90<\/h2>\n<p>\u4eba\u5728\u73af\u7684\u6982\u5ff5\u8d77\u6e90\u4e8e\u63a7\u5236\u5de5\u7a0b\uff0c\u8be5\u672f\u8bed\u7528\u4e8e\u63cf\u8ff0\u9700\u8981\u4eba\u7c7b\u4ea4\u4e92\u624d\u80fd\u6210\u529f\u8fd0\u884c\u7684\u7cfb\u7edf\u3002\u5b83\u7684\u7b2c\u4e00\u6b21\u91cd\u8981\u63d0\u53ca\u53ef\u4ee5\u8ffd\u6eaf\u5230 20 \u4e16\u7eaa 40 \u5e74\u4ee3\uff0c\u5f53\u65f6\u63a7\u5236\u8bba\u5174\u8d77\uff0c\u8be5\u9886\u57df\u7814\u7a76\u673a\u5668\u548c\u751f\u7269\u4f53\u56fa\u6709\u7684\u901a\u4fe1\u548c\u63a7\u5236\u7cfb\u7edf\u3002<\/p>\n<p>\u7136\u800c\uff0c\u968f\u7740\u6280\u672f\u8fdb\u6b65\u8bc1\u660e\u4e86\u4eba\u7c7b\u8ba4\u77e5\u80fd\u529b\u4e0e\u673a\u5668\u9a71\u52a8\u64cd\u4f5c\u76f8\u7ed3\u5408\u7684\u6f5c\u529b\uff0cHITL \u5728\u4eba\u5de5\u667a\u80fd\u9886\u57df\u7684\u5168\u9762\u5e94\u7528\u5728 21 \u4e16\u7eaa\u521d\u5f00\u59cb\u53d1\u5c55\u3002<\/p>\n<h2>\u63ed\u5f00\u201c\u4eba\u5728\u73af\u201d\u7684\u9762\u7eb1<\/h2>\n<p>\u4eba\u673a\u5faa\u73af\u7684\u6838\u5fc3\u662f\u4e00\u79cd\u673a\u5668\u5b66\u4e60\u65b9\u6cd5\uff0c\u4eba\u7c7b\u79ef\u6781\u53c2\u4e0e\u673a\u5668\u5b66\u4e60\u6a21\u578b\u751f\u547d\u5468\u671f\u7684\u4e0d\u540c\u9636\u6bb5\u3002\u4ece\u6570\u636e\u9884\u5904\u7406\u3001\u7279\u5f81\u63d0\u53d6\u3001\u6a21\u578b\u8bad\u7ec3\u5230\u6d4b\u8bd5\u548c\u90e8\u7f72\u540e\u53cd\u9988\uff0c\u4eba\u5de5\u5e72\u9884\u589e\u5f3a\u4e86\u4eba\u5de5\u667a\u80fd\u7cfb\u7edf\u7684\u529f\u80fd\u3002<\/p>\n<p>HITL \u4ece\u6839\u672c\u4e0a\u5efa\u7acb\u5728\u8fd9\u6837\u7684\u7406\u5ff5\u4e4b\u4e0a\uff1a\u867d\u7136\u4eba\u5de5\u667a\u80fd\u53ef\u4ee5\u8f7b\u677e\u5904\u7406\u91cd\u590d\u6027\u548c\u8ba1\u7b97\u5bc6\u96c6\u578b\u4efb\u52a1\uff0c\u4f46\u4eba\u7c7b\u5374\u5e26\u6765\u4e86\u4eba\u5de5\u667a\u80fd\u96be\u4ee5\u6a21\u4eff\u7684\u72ec\u7279\u5c5e\u6027\uff0c\u4f8b\u5982\u521b\u9020\u529b\u3001\u60c5\u5883\u7406\u89e3\u548c\u76f4\u89c9\u3002<\/p>\n<h2>\u4eba\u5728\u73af\u7684\u8fd0\u4f5c<\/h2>\n<p>HITL \u7cfb\u7edf\u901a\u8fc7\u534f\u4f5c\u6846\u67b6\u8fd0\u884c\uff0c\u5728\u8be5\u6846\u67b6\u4e2d\uff0c\u4eba\u548c\u673a\u5668\u90fd\u4e3a\u89e3\u51b3\u95ee\u9898\u7684\u8fc7\u7a0b\u505a\u51fa\u8d21\u732e\u3002\u4ee5\u4e0b\u662f\u5176\u529f\u80fd\u7684\u7b80\u5316\u89c6\u56fe\uff1a<\/p>\n<ol>\n<li><strong>\u9884\u5904\u7406\uff1a<\/strong> \u4eba\u7c7b\u7684\u53c2\u4e0e\u786e\u4fdd\u4e86\u6570\u636e\u96c6\u7684\u8d28\u91cf\u548c\u76f8\u5173\u6027\uff0c\u5305\u62ec\u6807\u7b7e\u548c\u6ce8\u91ca\u3002<\/li>\n<li><strong>\u8bad\u7ec3\uff1a<\/strong> \u6e05\u7406\u5e76\u6807\u8bb0\u7684\u6570\u636e\u96c6\u7528\u4e8e\u8bad\u7ec3 ML \u6a21\u578b\u3002<\/li>\n<li><strong>\u63a8\u7406\uff1a<\/strong> \u7ecf\u8fc7\u8bad\u7ec3\u7684\u6a21\u578b\u6839\u636e\u8f93\u5165\u8fdb\u884c\u9884\u6d4b\u3002<\/li>\n<li><strong>\u5ba1\u67e5\uff1a<\/strong> \u5982\u6709\u5fc5\u8981\uff0c\u4eba\u7c7b\u4f1a\u5ba1\u67e5\u5e76\u7ea0\u6b63\u6a21\u578b\u7684\u8f93\u51fa\u3002<\/li>\n<li><strong>\u53cd\u9988\uff1a<\/strong> \u6821\u6b63\u540e\u7684\u8f93\u51fa\u4f1a\u53cd\u9988\u5230\u7cfb\u7edf\u4e2d\uff0c\u4ece\u800c\u63d0\u9ad8\u6a21\u578b\u7684\u672a\u6765\u6027\u80fd\u3002<\/li>\n<\/ol>\n<p>\u6b64\u53cd\u9988\u5faa\u73af\u6301\u7eed\u8fdb\u884c\uff0c\u76f4\u5230\u6a21\u578b\u7684\u9884\u6d4b\u8fbe\u5230\u6240\u9700\u7684\u51c6\u786e\u5ea6\u3002<\/p>\n<h2>\u4eba\u5728\u73af\u7684\u4e3b\u8981\u7279\u70b9<\/h2>\n<p>\u4eba\u673a\u4ea4\u4e92\u4f5c\u4e3a\u4e00\u79cd\u6982\u5ff5\u548c\u5b9e\u8df5\uff0c\u5177\u6709\u51e0\u4e2a\u663e\u8457\u7684\u7279\u70b9\uff1a<\/p>\n<ul>\n<li><strong>\u534f\u4f5c\u667a\u80fd\uff1a<\/strong> HITL \u5c06\u673a\u5668\u7684\u8ba1\u7b97\u80fd\u529b\u4e0e\u4eba\u7c7b\u7684\u8ba4\u77e5\u6280\u80fd\u7ed3\u5408\u8d77\u6765\u3002<\/li>\n<li><strong>\u4e92\u52a8\u5b66\u4e60\uff1a<\/strong> \u8be5\u7cfb\u7edf\u4e0d\u65ad\u4ece\u4eba\u7c7b\u53cd\u9988\u4e2d\u5b66\u4e60\uff0c\u968f\u7740\u65f6\u95f4\u7684\u63a8\u79fb\u4e0d\u65ad\u63d0\u9ad8\u5176\u6027\u80fd\u3002<\/li>\n<li><strong>\u63d0\u9ad8\u51c6\u786e\u6027\uff1a<\/strong> \u4eba\u4e3a\u5e72\u9884\u6709\u52a9\u4e8e\u51cf\u5c11\u4eba\u5de5\u667a\u80fd\u7cfb\u7edf\u672c\u8eab\u53ef\u80fd\u72af\u7684\u9519\u8bef\u3002<\/li>\n<li><strong>\u591a\u529f\u80fd\u6027\uff1a<\/strong> HITL \u53ef\u5e94\u7528\u4e8e\u4ece\u81ea\u52a8\u9a7e\u9a76\u6c7d\u8f66\u5230\u533b\u7597\u8bca\u65ad\u7b49\u5e7f\u6cdb\u9886\u57df\u3002<\/li>\n<li><strong>\u4fe1\u4efb\u4e0e\u900f\u660e\u5ea6\uff1a<\/strong> \u901a\u8fc7\u8ba9\u4eba\u7c7b\u53c2\u4e0e\u51b3\u7b56\u8fc7\u7a0b\uff0cHITL \u63d0\u9ad8\u4e86\u4eba\u5de5\u667a\u80fd\u7cfb\u7edf\u7684\u900f\u660e\u5ea6\u548c\u4fe1\u4efb\u5ea6\u3002<\/li>\n<\/ul>\n<h2>\u4eba\u5728\u73af\u7cfb\u7edf\u7684\u7c7b\u578b<\/h2>\n<p>HITL \u7cfb\u7edf\u6709\u591a\u79cd\u7c7b\u578b\uff0c\u6839\u636e\u4eba\u4e3a\u5e72\u9884\u7684\u6c34\u5e73\u548c\u6027\u8d28\u8fdb\u884c\u5206\u7c7b\uff1a<\/p>\n<table>\n<thead>\n<tr>\n<th><strong>\u7c7b\u578b<\/strong><\/th>\n<th><strong>\u63cf\u8ff0<\/strong><\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td><strong>\u88ab\u52a8HITL<\/strong><\/td>\n<td>\u4eba\u5de5\u8f93\u5165\u4ec5\u7528\u4e8e\u521d\u59cb\u8bad\u7ec3\u6216\u5b9a\u671f\u66f4\u65b0\u3002<\/td>\n<\/tr>\n<tr>\n<td><strong>\u4e3b\u52a8HITL<\/strong><\/td>\n<td>\u4eba\u7c7b\u4e0d\u65ad\u53c2\u4e0e\uff0c\u5b9e\u65f6\u9a8c\u8bc1\u548c\u7ea0\u6b63\u4eba\u5de5\u667a\u80fd\u7684\u9884\u6d4b\u3002<\/td>\n<\/tr>\n<tr>\n<td><strong>\u6df7\u5408HITL<\/strong><\/td>\n<td>\u88ab\u52a8\u548c\u4e3b\u52a8\u7684\u7ed3\u5408\uff0c\u4eba\u7c7b\u53c2\u4e0e\u521d\u59cb\u8bad\u7ec3\u5e76\u5728\u4e0d\u786e\u5b9a\u7684\u60c5\u51b5\u4e0b\u88ab\u53ec\u5524\u3002<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<h2>\u5229\u7528\u4eba\u5728\u73af\uff1a\u6311\u6218\u548c\u89e3\u51b3\u65b9\u6848<\/h2>\n<p>HITL \u5728\u533b\u7597\u4fdd\u5065\u3001\u81ea\u52a8\u9a7e\u9a76\u6c7d\u8f66\u3001\u822a\u7a7a\u822a\u5929\u3001\u5ba2\u6237\u670d\u52a1\u7b49\u4f17\u591a\u9886\u57df\u90fd\u6709\u5e94\u7528\u3002\u7136\u800c\uff0c\u8fd9\u5e76\u975e\u6ca1\u6709\u6311\u6218\u3002\u53ef\u80fd\u5b58\u5728\u4e0e\u4eba\u7c7b\u53c2\u4e0e\u7684\u53ef\u6269\u5c55\u6027\u3001\u6570\u636e\u9690\u79c1\u4ee5\u53ca\u4eba\u7c7b\u53cd\u9988\u4e2d\u7684\u6f5c\u5728\u504f\u5dee\u76f8\u5173\u7684\u95ee\u9898\u3002<\/p>\n<p>\u5c3d\u7ba1\u5982\u6b64\uff0c\u8fd9\u4e9b\u6311\u6218\u662f\u53ef\u4ee5\u7f13\u89e3\u7684\u3002\u5bf9\u4e8e\u53ef\u6269\u5c55\u6027\u800c\u8a00\uff0c\u4e3b\u52a8\u5b66\u4e60\u7b49\u6280\u672f\u53ef\u4ee5\u901a\u8fc7\u4ec5\u5728\u5fc5\u8981\u65f6\u53c2\u4e0e\u6765\u5e2e\u52a9\u51cf\u5c11\u4eba\u529b\u3002\u53ef\u4ee5\u901a\u8fc7\u533f\u540d\u5316\u4e2a\u4eba\u6570\u636e\u548c\u5b9e\u65bd\u4e25\u683c\u7684\u6570\u636e\u6cbb\u7406\u5b9e\u8df5\u6765\u7ef4\u62a4\u9690\u79c1\u3002\u6700\u540e\uff0c\u4e3a\u4e86\u7ba1\u7406\u504f\u89c1\uff0c\u53ef\u4ee5\u96c7\u7528\u4e0d\u540c\u7684\u4eba\u7c7b\u5ba1\u9605\u8005\u7fa4\u4f53\u3002<\/p>\n<h2>\u5c06\u4eba\u5728\u73af\u4e0e\u7c7b\u4f3c\u6982\u5ff5\u8fdb\u884c\u6bd4\u8f83<\/h2>\n<p>\u4e0b\u8868\u5c06 HITL \u4e0e\u7c7b\u4f3c\u672f\u8bed\u8fdb\u884c\u4e86\u6bd4\u8f83\uff1a<\/p>\n<table>\n<thead>\n<tr>\n<th><strong>\u6982\u5ff5<\/strong><\/th>\n<th><strong>\u63cf\u8ff0<\/strong><\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td><strong>\u4eba\u673a\u4ea4\u4e92<\/strong><\/td>\n<td>\u5728\u673a\u5668\u5b66\u4e60\u6a21\u578b\u7684\u6574\u4e2a\u751f\u547d\u5468\u671f\u4e2d\u6d89\u53ca\u4eba\u7c7b\u53cd\u9988\u3002<\/td>\n<\/tr>\n<tr>\n<td><strong>\u4eba\u5728\u73af\u8def<\/strong><\/td>\n<td>\u4eba\u7c7b\u76d1\u7763\u4eba\u5de5\u667a\u80fd\u7684\u8fd0\u884c\u5e76\u4ec5\u5728\u5fc5\u8981\u65f6\u8fdb\u884c\u5e72\u9884\u3002<\/td>\n<\/tr>\n<tr>\n<td><strong>\u4eba\u4e3a\u8131\u79bb\u5faa\u73af<\/strong><\/td>\n<td>\u4eba\u5de5\u667a\u80fd\u5b8c\u5168\u72ec\u7acb\u8fd0\u884c\uff0c\u65e0\u9700\u4eba\u5de5\u5e72\u9884\u3002<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<h2>\u4eba\u5728\u73af\u7684\u672a\u6765\u5c55\u671b<\/h2>\n<p>HITL \u7684\u672a\u6765\u4f3c\u4e4e\u5145\u6ee1\u5e0c\u671b\uff0c\u6f5c\u5728\u7684\u8fdb\u6b65\u96c6\u4e2d\u5728\u4eba\u7c7b\u8ba4\u77e5\u4e0e\u4eba\u5de5\u667a\u80fd\u7684\u66f4\u6df1\u5c42\u6b21\u6574\u5408\u4e0a\u3002\u8111\u673a\u63a5\u53e3\u548c\u60c5\u611f\u8ba1\u7b97\u7b49\u6280\u672f\u53ef\u80fd\u662f\u5173\u952e\u8d21\u732e\u8005\u3002\u8fd9\u4e2a\u60f3\u6cd5\u662f\u8ba9\u4eba\u5de5\u667a\u80fd\u66f4\u5177\u540c\u7406\u5fc3\u3001\u9053\u5fb7\u548c\u9002\u5e94\u6027\uff0c\u4fc3\u8fdb\u4eba\u7c7b\u548c\u4eba\u5de5\u667a\u80fd\u4e4b\u95f4\u7684\u65e0\u7f1d\u534f\u4f5c\u3002<\/p>\n<h2>\u4ee3\u7406\u670d\u52a1\u5668\u548c\u4eba\u5728\u73af<\/h2>\n<p>\u4ee3\u7406\u670d\u52a1\u5668\uff08\u4f8b\u5982 OneProxy \u63d0\u4f9b\u7684\u4ee3\u7406\u670d\u52a1\u5668\uff09\u53ef\u4ee5\u5728 HITL \u7cfb\u7edf\u4e2d\u53d1\u6325\u91cd\u8981\u4f5c\u7528\u3002\u5b83\u4eec\u53ef\u4ee5\u4e3a\u6b63\u5728\u4f7f\u7528\u7684\u6570\u636e\u63d0\u4f9b\u4e00\u5c42\u5b89\u5168\u4fdd\u62a4\uff0c\u786e\u4fdd\u9690\u79c1\u548c\u5408\u89c4\u6027\u3002\u6b64\u5916\uff0c\u5b83\u4eec\u8fd8\u53ef\u7528\u4e8e\u4e3a\u673a\u5668\u5b66\u4e60\u6a21\u578b\u521b\u5efa\u66f4\u771f\u5b9e\u3001\u66f4\u591a\u6837\u5316\u7684\u6d4b\u8bd5\u73af\u5883\u3002\u8fd9\u53ef\u4ee5\u663e\u7740\u63d0\u9ad8\u6a21\u578b\u7684\u9c81\u68d2\u6027\u548c\u901a\u7528\u6027\u3002<\/p>\n<h2>\u76f8\u5173\u94fe\u63a5<\/h2>\n<ol>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Human-in-the-loop\" target=\"_new\" rel=\"noopener nofollow\">\u4eba\u5728\u73af\u673a\u5668\u5b66\u4e60<\/a><\/li>\n<li><a href=\"https:\/\/arxiv.org\/abs\/2006.12461\" target=\"_new\" rel=\"noopener nofollow\">\u4eba\u5728\u73af\uff0c\u4eba\u5de5\u667a\u80fd\u4f26\u7406\u54f2\u5b66<\/a><\/li>\n<li><a href=\"https:\/\/databricks.com\/glossary\/human-in-the-loop-for-machine-learning\" target=\"_new\" rel=\"noopener nofollow\">\u673a\u5668\u5b66\u4e60\u7684\u4eba\u673a\u4ea4\u4e92<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Proxy_server\" target=\"_new\" rel=\"noopener nofollow\">\u4ee3\u7406\u670d\u52a1\u5668<\/a><\/li>\n<\/ol>","protected":false},"featured_media":468577,"menu_order":0,"template":"","meta":{"_acf_changed":false,"content-type":"","inline_featured_image":false,"footnotes":""},"class_list":["post-477510","wiki","type-wiki","status-publish","has-post-thumbnail","hentry"],"acf":{"faq_title":"Frequently Asked Questions about <mark>Human-in-the-Loop: An Insight into Collaborative Computing<\/mark>","faq_items":[{"question":"What is Human-in-the-Loop (HITL)?","answer":"<p>Human-in-the-Loop is an interactive approach to computing that integrates human intelligence and input into the artificial intelligence (AI) systems' workflow. It's about using human insights at different stages of the machine learning model's life cycle, including data pre-processing, feature extraction, model training, testing, and post-deployment feedback.<\/p>"},{"question":"What is the history of Human-in-the-Loop?","answer":"<p>The concept of Human-in-the-Loop originated in control engineering, where systems required human interaction for operation. The first significant mention dates back to the 1940s in the field of cybernetics. The application of HITL in artificial intelligence, however, began to evolve in the early 21st century with advancements in technology.<\/p>"},{"question":"How does a Human-in-the-Loop system work?","answer":"<p>A HITL system functions through a collaborative framework involving humans and machines. It starts with humans pre-processing data, followed by the machine training on this data. The model then makes predictions, which humans review and correct, if necessary. These corrected outputs are then fed back into the system, which learns and improves from this feedback. This loop continues until the model's predictions reach a satisfactory level of accuracy.<\/p>"},{"question":"What are the key features of Human-in-the-Loop?","answer":"<p>The key features of HITL include collaborative intelligence, interactive learning, improved accuracy, versatility across various domains, and enhanced trust and transparency in AI systems.<\/p>"},{"question":"What types of Human-in-the-Loop systems exist?","answer":"<p>HITL systems can be categorized into Passive HITL, where human input is used for initial training or periodic updates; Active HITL, where humans continually validate and correct AI predictions; and Hybrid HITL, which combines the elements of both passive and active types.<\/p>"},{"question":"What are some challenges and solutions related to the use of Human-in-the-Loop?","answer":"<p>Challenges related to the use of HITL include scalability of human involvement, data privacy, and potential biases in human feedback. These can be addressed by using active learning techniques, implementing data anonymization and robust governance practices, and employing a diverse group of human reviewers to manage biases.<\/p>"},{"question":"How do proxy servers relate to Human-in-the-Loop?","answer":"<p>Proxy servers, such as those provided by OneProxy, can offer security for data used in HITL systems, ensuring privacy and compliance. They can also be used to create diverse and realistic testing environments for machine learning models, thus improving their robustness and generalizability.<\/p>"},{"question":"What are the future perspectives of Human-in-the-Loop?","answer":"<p>Future perspectives of HITL include deeper integration of human cognition with AI. Potential advancements could focus on technologies like brain-computer interfaces and affective computing, with an aim to make AI systems more empathetic, ethical, and adaptable.<\/p>"}]},"_links":{"self":[{"href":"https:\/\/oneproxy.pro\/cn\/wp-json\/wp\/v2\/wiki\/477510","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/oneproxy.pro\/cn\/wp-json\/wp\/v2\/wiki"}],"about":[{"href":"https:\/\/oneproxy.pro\/cn\/wp-json\/wp\/v2\/types\/wiki"}],"version-history":[{"count":0,"href":"https:\/\/oneproxy.pro\/cn\/wp-json\/wp\/v2\/wiki\/477510\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/oneproxy.pro\/cn\/wp-json\/wp\/v2\/media\/468577"}],"wp:attachment":[{"href":"https:\/\/oneproxy.pro\/cn\/wp-json\/wp\/v2\/media?parent=477510"}],"curies":[{"name":"\u53ef\u6e7f\u6027\u7c89\u5242","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}