| { | |
| "paper_id": "O16-1032", | |
| "header": { | |
| "generated_with": "S2ORC 1.0.0", | |
| "date_generated": "2023-01-19T08:04:53.622936Z" | |
| }, | |
| "title": "Multi-Channel Television Echo Cancellation based on Deep Recurrent Neural Networks", | |
| "authors": [ | |
| { | |
| "first": "\u9ec3\u5b8f", | |
| "middle": [], | |
| "last": "Huang", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "", | |
| "institution": "National Taipei University of Technology", | |
| "location": {} | |
| }, | |
| "email": "" | |
| }, | |
| { | |
| "first": "Hung", | |
| "middle": [], | |
| "last": "\u570b\u7acb\u53f0\u5317\u79d1\u6280\u5927\u5b78\u96fb\u5b50\u5de5\u7a0b\u5b78\u7cfb", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "", | |
| "institution": "National Taipei University of Technology", | |
| "location": {} | |
| }, | |
| "email": "" | |
| }, | |
| { | |
| "first": "\u6d2a\u744b\u5db8", | |
| "middle": [], | |
| "last": "Hung", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "", | |
| "institution": "National Taipei University of Technology", | |
| "location": {} | |
| }, | |
| "email": "" | |
| }, | |
| { | |
| "first": "Wei-Jung", | |
| "middle": [], | |
| "last": "\u570b\u7acb\u53f0\u5317\u79d1\u6280\u5927\u5b78\u96fb\u5b50\u5de5\u7a0b\u5b78\u7cfb", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "", | |
| "institution": "National Taipei University of Technology", | |
| "location": {} | |
| }, | |
| "email": "" | |
| }, | |
| { | |
| "first": "\u5ed6\u5143\u752b", | |
| "middle": [], | |
| "last": "Liao", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "", | |
| "institution": "National Taipei University of Technology", | |
| "location": {} | |
| }, | |
| "email": "[email protected]" | |
| }, | |
| { | |
| "first": "Yuan-Fu", | |
| "middle": [], | |
| "last": "\u570b\u7acb\u53f0\u5317\u79d1\u6280\u5927\u5b78\u96fb\u5b50\u5de5\u7a0b\u5b78\u7cfb", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "", | |
| "institution": "National Taipei University of Technology", | |
| "location": {} | |
| }, | |
| "email": "" | |
| } | |
| ], | |
| "year": "", | |
| "venue": null, | |
| "identifiers": {}, | |
| "abstract": "", | |
| "pdf_parse": { | |
| "paper_id": "O16-1032", | |
| "_pdf_hash": "", | |
| "abstract": [], | |
| "body_text": [], | |
| "back_matter": [], | |
| "bib_entries": { | |
| "BIBREF1": { | |
| "ref_id": "b1", | |
| "title": "Nonliner Acoustic Echo Cancellstion With 2nd Order Adaptive Volterra Filters", | |
| "authors": [ | |
| { | |
| "first": "A", | |
| "middle": [], | |
| "last": "Stenger", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "L", | |
| "middle": [], | |
| "last": "Trautmann", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "R", | |
| "middle": [], | |
| "last": "Rabenstein", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1999, | |
| "venue": "IEEE Int. Conf. on Acoustics, Speech & Signal Procrssing(ICASSP)", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "A. Stenger, L. Trautmann and R. Rabenstein, \"Nonliner Acoustic Echo Cancellstion With 2nd Order Adaptive Volterra Filters,\" IEEE Int. Conf. on Acoustics, Speech & Signal Procrssing(ICASSP), 1999.", | |
| "links": null | |
| }, | |
| "BIBREF2": { | |
| "ref_id": "b2", | |
| "title": "Adaptive Filtering Algorithms and Practical Implementation 4 th", | |
| "authors": [ | |
| { | |
| "first": "S", | |
| "middle": [ | |
| "R" | |
| ], | |
| "last": "Paulo", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "", | |
| "middle": [], | |
| "last": "Diniz", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2012, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "469--477", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Paulo S. R. Diniz, Adaptive Filtering Algorithms and Practical Implementation 4 th ,New York\uff1aSpringer Verlag,2012,pp.469-477.", | |
| "links": null | |
| }, | |
| "BIBREF5": { | |
| "ref_id": "b5", | |
| "title": "Implementation of BP and RBF neural network and their performance comparison", | |
| "authors": [ | |
| { | |
| "first": "Liu", | |
| "middle": [], | |
| "last": "Yong", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Zhang", | |
| "middle": [], | |
| "last": "Liyi", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2007, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Liu Yong and Zhang Liyi, Implementation of BP and RBF neural network and their performance comparison, Master Thesis, University of Technology, 2007.", | |
| "links": null | |
| } | |
| }, | |
| "ref_entries": { | |
| "TABREF0": { | |
| "num": null, | |
| "content": "<table><tr><td>\u4e00\u3001\u7c21\u4ecb \u9593\u592a\u4e45\uff0c\u5176\u6548\u679c\u901a\u5e38\u4e0d\u597d\u3002 \u53e6\u5916\uff0c\u8f3b\u5c04\u57fa\u5e95\u51fd\u6578\u985e\u795e\u7d93\u7db2\u8def(RBFNN)\u5177\u6709\u903c\u8fd1\u4efb\u610f\u7684\u975e\u7dda\u6027\u51fd\u6578\u7684\u80fd\u529b\uff0c\u800c\u4e14\u5177\u5099 \u5c64\u7684\u905e\u8ff4\u985e\u795e\u7d93\u7db2\u8def\u5c31\u8db3\u4ee5\u7372\u53d6\u7684\u9577\u6642\u9593\u8cc7\u8a0a\uff0c\u56e0\u6b64\u591a\u5c64\u905e\u56de\u985e\u795e\u7d93\u7db2\u8def\u5c31\u80fd\u770b\u5f97\u66f4 \u5176\u4e2d x(n)\u539f\u59cb\u96fb\u8996\u7bc0\u76ee\u8072\uff0c\u4e26\u628a\u5b83\u7576\u6210\u8072\u5b78\u56de\u8072\u7cfb\u7d71\u7684\u53c3\u8003\u4fe1\u865f\uff0c\u540c\u6642\u4e5f\u662f\u56db\u500b\u6df1\u5c64\u985e (\u4e00) \u8a9e\u6599\u8aaa\u660e \u56de\u8072\u6d88\u9664\u5f8c\u5f97\u5230\u7684 TCC300 \u6e2c\u8a66\u4eba\u8072 s(n)\u5169\u8005\u76f8\u4e92\u53bb\u6bd4\u8f03\uff0c\u7576\u5206\u5b50 s(n)\u8d8a\u5c0f\u6642\uff0c\u6b64\u6642 2. \u5be6\u9a57\u4e8c\uff0c\u4eba\u8072\u6df7\u96fb\u8996\u7bc0\u76ee\u8072\u5be6\u9a57: \u7d93\u7531\u9019\u4e9b\u6a21\u64ec\u5be6\u9a57\u7d50\u679c\u53ef\u4ee5\u767c\u73fe\uff0c\u5728(A)\u55ae\u901a\u9053\u5be6\u9a57\u4e00\u7d14\u96fb\u8996\u7bc0\u76ee\u8072\u6d88\u9664\u4e0a\uff0c\u5c0d\u65bc\u67d0\u4e9b</td></tr><tr><td>\u8072\u63a7\u96fb\u8996\u662f\u975e\u5e38\u4eba\u6027\u5316\u7684\u529f\u80fd\uff0c\u4f46\u7919\u65bc\u96fb\u8996\u56de\u8072\u8207\u96dc\u8a0a\u7b49\u554f\u984c\u5f71\u97ff\uff0c\u5e38\u6703\u5e72\u64fe\u4f7f\u7528\u8005\u8a9e \u4e00\u822c\u5316\u80fd\u529b\uff0c\u5c0d\u672a\u77e5\u8cc7\u6599\u80fd\u6709\u6548\u5730\u8655\u7406\uff0c\u518d\u52a0\u4e0a\u5feb\u901f\u7684\u5b78\u7fd2\u6536\u6582\u901f\u5ea6\u53ca\u4f4e\u8a08\u7b97\u8907\u96dc\u5ea6\uff0c \u5ee3\uff0c\u6293\u53d6\u66f4\u5927\u7684\u8cc7\u6599\u3002\u4ee5\u4e0b\u5148\u4ee5\u55ae\u8072\u9053\u56de\u8072\u6d88\u9664\u7684\u505a\u6cd5\uff0c\u7136\u5f8c\u5728\u8aaa\u660e\u591a\u8072\u9053\u8ff4\u97f3\u6d88\u9664\u7684 \u795e\u7d93\u7db2\u8def\u7684\u8f38\u5165\u3002i \u70ba\u9ea5\u514b\u98a8\u8072\u9053\u6578\uff0ci=1,2,3,4\u3002mi(n)\u70ba Kinect for Xbox one \u9ea5\u514b\u98a8\u9663\u5217 \u6e2c\u8a66\u4eba\u8072\u9304\u97f3\u8a9e\u6599\u5f9e TCC300 \u8a9e\u6599\u5eab\u4e2d\u9078\u64c7\u4e86 4 \u7537 4 \u5973\u7684\u97f3\u6a94\uff0c\u4e14\u70ba\u96a8\u6a5f\u64f7\u53d6\u5341\u79d2\u9418\u7247 ERLE \u503c\u5c31\u6108\u5927\uff0c\u4ee3\u8868\u6d88\u9664\u6027\u80fd\u6108\u597d\uff1b\u4e5f\u8868\u793a\u5f97\u5230\u6108\u6e05\u6670\u7684 TCC300 \u6e2c\u8a66\u4eba\u8072\u3002 \u4ee5\u4e0b\u662f\u516b\u500b\u4e0d\u540c\u7684\u6e2c\u8a66\u4f7f\u7528\u8005\u6df7\u96fb\u8996\u7bc0\u76ee\u8072\u7684\u56de\u8072\u6d88\u9664\u5be6\u9a57\uff0c\u53ef\u85c9\u6b64\u89c0\u6e2c\u4e0d\u540c\u6e2c\u8a66\u4f7f\u7528 \u7a2e\u985e\u7684\u7bc0\u76ee\u800c\u8a00\uff0c\u6709\u4e9b\u6f14\u7b97\u6cd5\u6d88\u9664\u6548\u80fd\u8868\u73fe\u6bd4\u8f03\u7a81\u51fa\uff0c\u4f8b\u5982\uff1aNLMS \u5728\u97f3\u6a02\u985e\u3001RBF</td></tr><tr><td>\u97f3\u64cd\u4f5c\u3002\u56e0\u6b64\u4e00\u822c\u9700\u8981\u52a0\u4e0a\u8072\u5b78\u56de\u8072\u6d88\u9664\u7cfb\u7d71\uff0c\u4ee5\u9069\u61c9\u6027\u6ffe\u6ce2\u5668\u6f14\u7b97\u6cd5[1-2]\uff0c\u5728\u5be6\u969b\u7a7a \u4f46\u5728\u89c0\u770b\u96fb\u8996\u6642\uff0c\u56e0\u96fb\u8996\u7bc0\u76ee\u8072\u97f3\u901a\u5e38\u958b\u5f88\u5927\u8072\uff0c\u4e14\u56de\u8072\u7d93\u904e\u91cd\u91cd\u53cd\u5c04\uff0c\u6b98\u97ff\u6642\u9593\u901a\u5e38 \u5df2\u6210\u529f\u5ee3\u6cdb\u61c9\u7528\u5728\u975e\u7dda\u6027\u56de\u8072\u6d88\u9664\uff0c\u4e0b\u5716 3 \u70ba RBF \u985e\u795e\u7d93\u7db2\u8def\u7528\u65bc\u56de\u8072\u6d88\u9664\u7684\u67b6\u69cb\u5716\u3002 \u505a\u6cd5\u3002 \u6240\u9304\u88fd\u7684\u96fb\u8996\u7bc0\u76ee\u8072\u53ca\u4f7f\u7528\u8005\u8aaa\u8a71\u8072\u7684\u6df7\u5408\u8072\u97f3\u3002x\u0302i(n)\u70ba\u96fb\u8996\u7bc0\u76ee\u8072\u7684\u56de\u8072\uff0ci=1,2,3,4\u3002 \u6bb5\u8aaa\u8a71\u8072\uff1b\u800c\u6240\u56de\u9304\u7684\u96fb\u8996\u7bc0\u76ee\u8072\u70ba\u56db\u5927\u985e\uff0c\u6bcf\u985e\u70ba\u6709 10 \u500b\u97f3\u6a94\uff0c\u5f9e\u4e2d\u96a8\u6a5f\u64f7\u53d6\u5341\u79d2 (\u56db) \u5be6\u9a57\u7d50\u679c \u8005\u5728\u6bcf\u4e00\u985e\u96fb\u8996\u7bc0\u76ee\u8072\u7684\u6f14\u7b97\u6cd5\u5e73\u5747\u6548\u80fd\uff0c\u6700\u5f8c\u518d\u770b\u516b\u500b\u4eba\u7684\u7e3d\u5e73\u5747\uff0c\u7d50\u679c\u5982\u4e0b\u5716 11 \u5728\u904b\u52d5\u985e\u7b49\uff1b\u4e5f\u986f\u793a\u4e86\u975e\u7dda\u6027\u6ffe\u6ce2\u6f14\u7b97\u6cd5\u5c0d\u65bc\u6bd4\u8f03\u8907\u96dc\u7684\u74b0\u5883\u5177\u6709\u8f03\u4f73\u7684\u6d88\u9664\u56de\u8072\u80fd</td></tr><tr><td>\u9593\u74b0\u5883\u4e0b\uff0c\u5b78\u7fd2\u56de\u8072\u8def\u5f91\uff0c\u9810\u6e2c\u4e26\u6d88\u9664\u96fb\u8996\u7bc0\u76ee\u56de\u8072\uff0c\u589e\u5f37\u4f7f\u7528\u8005\u8a9e\u97f3\u8cea\u91cf\u3002 \u5f88\u9577\u3002\u800c\u4e14\u96fb\u8996\u7bc0\u76ee\u662f\u52d5\u614b\u6301\u7e8c\u7684\u64ad\u653e\u8457\uff0c\u4e0a\u4e00\u500b\u6642\u9593\u9ede\u64ad\u653e\u7684\u8072\u97f3\u6703\u5f71\u97ff\u4e0b\u4e00\u500b\u6642\u9593 (\u4e00) \u55ae\u8072\u9053\u6df1\u5c64\u905e\u8ff4\u985e\u795e\u7d93\u7db2\u8def\u56de\u97f3\u6d88\u9664 s(n)\u70ba\u4f7f\u7528\u8005\u8aaa\u8a71\u8072\uff0cs\u0302i(n)\u70ba\u8aaa\u8a71\u7684\u56de\u8072\u3002x\u0129(n)\u70ba\u903c\u8fd1 x\u0302(n)\u7684\u4f30\u8a08\u503c\uff0c\u5373\u70ba\u56de\u8072\u62b5\u6d88\u9810\u6e2c \u9418\u7247\u6bb5\u7bc0\u76ee\u8072\uff0c\u4e14\u6bcf\u4e00\u985e\u6709 10 \u500b\u97f3\u6a94\uff1b\u6240\u4ee5\u5171\u6709 40 \u500b\u6e2c\u8a66\u80cc\u666f\u96fb\u8996\u7bc0\u76ee\u8072\u3002\u7531\u65bc TCC300 \u5728\u4ee5\u4e0b\u5be6\u9a57\u4e2d\uff0c\u6e2c\u8a66\u4eba\u8072\u8207\u96fb\u8996\u7bc0\u76ee\u8072\u6df7\u5408\u7684\u6bd4\u4f8b\u7686\u4fdd\u6301\u4e00\u6a23\u5927\u8072\uff0c\u5373\u6a21\u64ec SNR = 0dB \u6240\u793a\uff0c\u7576\u96fb\u8996\u56de\u8072\u6df7\u5408\u4eba\u8072\u6642\uff0c\u6240\u6709\u65b9\u6cd5\u90fd\u6703\u8b8a\u5dee\uff0c\u5176\u4e2d\u4ee5 RBF \u53d7\u5230\u7684\u5f71\u97ff\u6700\u5c0f\uff0cRNN \u529b\u3002\u800c\u5728\u5be6\u9a57\u4e8c\u4eba\u8072\u6df7\u96fb\u8996\u7bc0\u76ee\u8072\u6d88\u9664\u4e0a\uff0c\u53ef\u4ee5\u767c\u73fe\u5176\u4eba\u8072\u6703\u6ea2\u5165\u5230\u9069\u61c9\u6027\u6ffe\u6ce2\u5668\u4e2d\uff0c</td></tr><tr><td>\u9ede\u7684\u8072\u97f3\u3002\u82e5\u6211\u5011\u60f3\u6355\u6349\u9577\u671f\u6642\u9593\u95dc\u806f\u7684\u8cc7\u8a0a\uff0c\u5c31\u5fc5\u9808\u4f7f\u7528\u542b\u6709\u56de\u6388\u529f\u80fd\u7684\u6df1\u5c64\u905e\u8ff4\u985e \u4e0b\u5716 5 \u5176\u904b\u4f5c\u539f\u7406\u70ba\u904b\u7528\u9ea5\u514b\u98a8\u6240\u6536\u96c6\u5230\u7684\u8a9e\u97f3 m(n)\u8207\u900f\u904e\u96fb\u8996\u97f3\u6e90\u7dda\u8f38\u51fa\u53d6\u7684\u96fb\u8996 \u4fe1\u865f\u3002s\u0129(n)\u70ba mi(n)-x\u0129(n)\u7684\u8aa4\u5dee\u4fe1\u865f\uff1b\u5373\u6211\u5011\u60f3\u8981\u7372\u5f97\u7684\u4f7f\u7528\u8005\u8aaa\u8a71\u8072\u3002s\u0302out(n)\u70ba\u56db\u500b \u7684\u8a9e\u6599\u6a94\u6848\u683c\u5f0f\u70ba.pcm \u6a94\u8207\u6240\u4e0b\u8f09\u7684\u96fb\u8996\u7bc0\u76ee\u8072\u97f3\u58d3\u7e2e\u683c\u5f0f\u70ba MP4 \u6a94\uff0c\u52a0\u4e0a\u5169\u8005\u8a9e\u97f3 \u7684\u60c5\u5f62\u3002\u4ee5\u4e0b\u9032\u884c\u5169\u5927\u985e\u5be6\u9a57\uff0c\u5305\u522e(A)\u55ae\u8072\u9053(B)\u591a\u901a\u9053\u96fb\u8996\u56de\u8072\u6d88\u9664\u5be6\u9a57\u3002 \u6b21\u4e4b\uff0c\u4e14\u5169\u8005\u5dee\u8ddd\u4e0d\u5927\uff0c\u4f46\u6548\u679c\u90fd\u9060\u6bd4 NLMS \u8207 MLP \u8981\u597d\u3002 \u6240\u4ee5\u5b78\u7fd2\u51fa\u4f86\u7684\u96fb\u8996\u80cc\u666f\u8072\u6709\u4e9b\u8a31\u7684\u4eba\u8072\uff0c\u9019\u6703\u5f71\u97ff\u6d88\u9664\u6027\u80fd\uff0c\u4f46\u6574\u9ad4\u4f86\u8aaa\u975e\u7dda\u6027\u6ffe\u6ce2</td></tr><tr><td>\u8072\u5b78\u56de\u8072\u4e3b\u8981\u662f\u8072\u97f3\u7d93\u5587\u53ed\u64ad\u51fa\uff0c\u7531\u7a7a\u9593\u97ff\u61c9\u5c0e\u81f4\u7684\u4e00\u6b21\u6216\u591a\u6b21\u7684\u53cd\u5c04\u8072\u5230\u9ea5\u514b\u98a8\u6240\u5f15 \u795e\u7d93\u7db2\u8def[7]\uff0c\u5176\u628a\u4e0a\u500b\u6642\u9593\u9ede\u7684\u8f38\u51fa\u503c\u5b58\u4e0b\u4f86\uff0c\u4e26\u91cd\u65b0\u5c0e\u5165\u5230\u8f38\u5165\u7aef\uff0c\u4ee5\u4fbf\u5728\u6642\u9593\u4e0a \u7bc0\u76ee\u8072 x(n)\uff0c\u5229\u7528\u6df1\u5c64\u905e\u8ff4\u5f0f\u795e\u7d93\u7db2\u8def\u9810\u6e2c\u53ef\u80fd\u9304\u5230\u7684\u96fb\u8996\u56de\u8072 x(n)\uff0c\u76f8\u6e1b\u5f8c\u5c07\u56de\u8072\u6d88 RNN \u56de\u8072\u6d88\u9664\u7cfb\u7d71\u8f38\u51fa\u7684\u52a0\u6b0a\u5e73\u5747\u8aa4\u5dee\u4fe1\u865f\u3002 \u5167\u5bb9\u9577\u77ed\u5dee\u7570\u904e\u5927\uff0c\u5c0e\u81f4\u7121\u6cd5\u76f4\u63a5\u4f5c\u7a0b\u5f0f\u8f38\u5165\u6e2c\u8a66\u7684\u97f3\u6a94\uff0c\u6240\u4ee5\u5148\u4ee5\u97f3\u983b\u8655\u7406\u5de5\u5177\u4f5c\u683c \u6f14\u7b97\u6cd5\u5c0d\u65bc\u6df7\u8072\u6d88\u9664\u6548\u80fd\u8868\u73fe\u6bd4\u8f03\u4f73\u3002\u6700\u5f8c\u5728\u5be6\u9a57\u4e09\u900f\u904e\u5148\u5728\u524d\u4e94\u79d2\u7d14\u96fb\u8996\u7bc0\u76ee\u8072\u6642\u5b78</td></tr><tr><td>\u6458\u8981 \u672c\u8ad6\u6587\u7814\u7a76\u667a\u6167\u578b\u96fb\u8996\u64cd\u4f5c\u60c5\u5883\u4e0b\u4e4b\u96fb\u8996\u7bc0\u76ee\u56de\u8072\u6d88\u9664\uff0c\u5e0c\u671b\u80fd\u5728\u96fb\u8996\u7bc0\u76ee\u6301\u7e8c\u64ad \u653e\u7684\u60c5\u5f62\u4e0b\uff0c\u4ecd\u80fd\u9304\u5230\u8aaa\u8a71\u8005\u7684\u6e05\u6670\u8a9e\u97f3\uff0c\u4e26\u80fd\u61c9\u7528\u5728\u5373\u6642\u8a9e\u97f3\u901a\u8a0a\u8207\u9060\u8ddd\u8a9e\u97f3\u8fa8 \u8a8d\u4eba\u6a5f\u4ecb\u9762\u4e0a\u3002\u672c\u8ad6\u6587\u7684\u56de\u8072\u6d88\u9664\u7cfb\u7d71\u6f14\u7b97\u6cd5\u662f\u4ee5\u905e\u8ff4\u985e\u795e\u7d93\u7db2\u8def(Recurrent Neural Network\uff0cRNN)\u6f14\u7b97\u6cd5\uff0c\u518d\u914d\u4e0a\u591a\u901a\u9053\u9ea5\u514b\u98a8\u505a\u56de\u8072\u6d88\u9664\uff0c\u9054\u5230\u4eba\u8072\u589e\u5f37\uff0c \u6291\u5236\u566a\u97f3\u96dc\u8a0a\uff0c\u63d0\u9ad8\u8a9e\u97f3\u6e05\u6670\u5ea6\u3002\u5be6\u9a57\u5206\u5225\u5be6\u4f5c\u55ae\u7d14\u96fb\u8996\u7bc0\u76ee\u8072\u3001\u4eba\u8072\u6df7\u96fb\u8996\u7bc0\u76ee \u8072\u5169\u7a2e\u5be6\u9a57\uff0c\u518d\u5c0e\u5165\u524d\u4e94\u79d2\u7121\u4eba\u8072\u9810\u8a13\u7df4\uff0c\u5f8c\u4e94\u79d2\u6709\u4eba\u8072\u4e4b\u96fb\u8996\u7bc0\u76ee\u56de\u8072\u6d88\u9664\u6a21\u5f0f \u5be6\u9a57\uff0c\u5be6\u9a57\u7d50\u679c\u4ee5\u56de\u8072\u8870\u6e1b\u91cf\u4f86\u5224\u65b7\u6548\u80fd\u512a\u52a3\u3002\u5be6\u9a57\u986f\u793a\uff0c\u4ee5\u591a\u901a\u9053\u6df1\u5c64\u905e\u8ff4\u985e\u795e \u7d93\u7db2\u8def\u6548\u80fd\u512a\u65bc\u5176\u4ed6\u65b9\u6cd5\uff0c\u900f\u904e\u591a\u8072\u9053 RNN \u8655\u7406\uff0c\u7684\u78ba\u80fd\u6709\u6548\u5730\u6ffe\u9664\u96dc\u8a0a\u3002 \u95dc\u9375\u8a5e: \u8072\u5b78\u56de\u8072\u6d88\u9664\u3001\u9069\u61c9\u6027\u6ffe\u6ce2\u5668\u3001\u985e\u795e\u7d93\u7db2\u8def\u3001\u905e\u8ff4\u985e\u795e\u7d93\u7db2\u8def \u80fd\u5920\u6293\u53d6\u5927\u9577\u5ea6\u7684\u8f38\u5165\u8a0a\u865f\uff0c\u4f7f\u7cfb\u7d71\u80fd\u6709\u5927\u91cf\u7684\u6b77\u53f2\u8cc7\u6599\u53bb\u5b78\u7fd2\u96fb\u8996\u7bc0\u76ee\u56de\u8072\u7684\u8def\u5f91\u3002 \u9664\uff0c\u6b64\u6642\u7cfb\u7d71\u7684\u8aa4\u5dee\u8a0a\u865f s(n)\u70ba\u8f38\u51fa\u5f97\u5230\u7684\u6e05\u6670\u8a9e\u8005\u8072\u97f3\u3002 \u5f0f\u53ca\u6642\u9593\u4e0a\u7684\u6574\u7406\u3002\u4ee5\u4e0b\u8868 1 \u70ba\u8a9e\u6599\u683c\u5f0f\u6574\u7406\u3002 \u5716 8 90 \u5ea6\u89d2\u64fa\u8a2d\u5716 (A)\u55ae\u901a\u9053\u96fb\u8996\u56de\u8072\u6d88\u9664\u5be6\u9a57 \u7fd2\u80cc\u666f\u56de\u8072\u8def\u5f91\uff0c\u63a5\u8457\u5f8c\u4e94\u79d2\u4eba\u8072\u6df7\u96fb\u8996\u7bc0\u76ee\u8072\u76f4\u63a5\u7528\u8a13\u7df4\u5f8c\u7684\u7cfb\u7d71\u4f5c\u56de\u8072\u6d88\u9664\uff0c\u85c9\u6b64 \u8d77\uff0c\u4e3b\u6d41\u7684\u56de\u8072\u6d88\u9664\u65b9\u6cd5\u904b\u7528\u5176\u67b6\u69cb\u5982\u5716 1 \u6240\u793a\uff0c\u5176\u4f7f\u7528\u9069\u61c9\u6027\u6f14\u7b97\u6cd5\u81ea\u52d5\u5730\u8abf\u6574\u6ffe\u6ce2 \u5668\u6b0a\u91cd\u4fc2\u6578\uff0c\u4f7f\u8f38\u51fa\u4fe1\u865f\u80fd\u5920\u903c\u8fd1\u6240\u671f\u671b\u7684\u4fe1\u865f\u3002 \u5716 1 \u5178\u578b\u8072\u5b78\u56de\u8072\u6d88\u9664\u7cfb\u7d71\u67b6\u69cb\u5716 \u9069\u61c9\u6027\u6ffe\u6ce2\u5668\u6f14\u7b97\u6cd5\u6709\u7dda\u6027\u8207\u975e\u7dda\u6027\u6d88\u9664\u5169\u7a2e\u4f5c\u6cd5\uff0c\u50b3\u7d71\u4e0a\u7dda\u6027\u65b9\u6cd5\u5e38\u4ee5\u6b63\u898f\u5316\u6700\u5c0f\u5747 \u65b9\u6f14\u7b97\u6cd5(Normalized least mean squares\uff0cNLMS)[3-4]\u4f86\u5be6\u73fe\uff0c\u4f46\u662f\u7531\u65bc\u5728\u5ba4\u5167\u80cc\u666f\u96dc\u8a0a \u773e\u591a\uff0cNLMS \u6f14\u7b97\u6cd5\u672a\u80fd\u6709\u6548\u89e3\u6c7a\u975e\u7dda\u6027\u56de\u8072\uff0c\u65bc\u662f\u5e38\u9032\u4e00\u6b65\u5f15\u9032\u975e\u7dda\u6027\u9069\u61c9\u6027\u6f14\u7b97 \u6cd5\uff0c\u6539\u5584\u975e\u7dda\u6027\u56de\u8072\u7684\u90e8\u5206\uff0c\u5e38\u7528\u7684\u975e\u7dda\u6027\u6f14\u7b97\u6cd5\u5305\u62ec\u8f3b\u5c04\u57fa\u5e95\u51fd\u6578\u985e\u795e\u7d93\u7db2\u8def(Radial Basis Function Neural Network\uff0cRBFNN)[3,5-6]\u8207\u591a\u5c64\u611f\u77e5\u6a5f(Multi-layer Perceptron Neural Network\uff0cMLP )[3, 5-6]\u6f14\u7b97\u6cd5\uff0c\u5176\u900f\u904e\u5177\u9069\u61c9\u6027\u5b78\u7fd2\u80fd\u529b\u7684\u985e\u795e\u7d93\u7db2\u8def\u6f14\u7b97\u6cd5\uff0c \u514b\u670d\u975e\u7dda\u6027\u7684\u56e0\u7d20\u3002\u4e0d\u904e RBF \u8207 MLP \u96d6\u7136\u76f8\u7576\u6709\u6548\uff0c\u4f46\u56e0\u5176\u8f38\u5165\u8a0a\u865f\u7684\u9577\u5ea6\u901a\u5e38\u662f\u6709 \u9650\u7684\u4e14\u4e0d\u80fd\u592a\u9577\uff0c\u4ee5\u514d\u904b\u7b97\u592a\u4e45\uff0c\u56e0\u6b64\u901a\u5e38\u53ea\u80fd\u8655\u7406\u8f03\u77ed\u6642\u9593\u7684\u56de\u8072\uff0c\u82e5\u56de\u8072\u7684\u5f71\u97ff\u6642 \u6b64\u5916\u6211\u5011\u4e26\u8003\u616e\u7a7a\u9593\u8cc7\u8a0a\uff0c\u6539\u7528\u591a\u652f\u9ea5\u514b\u98a8\u7d44\u6210\u7684\u591a\u901a\u9053\u7cfb\u7d71[8]\uff0c\u4ee5\u5f37\u5316\u8a9e\u97f3\u8f38\u5165\u8a0a \u865f\u3002\u56e0\u6b64\uff0c\u57fa\u65bc\u4ee5\u4e0a\u5169\u9ede\u8003\u91cf\uff0c\u5728\u672c\u8ad6\u6587\u4e2d\u6211\u5011\u5c07\u63d0\u51fa\u591a\u901a\u9053\u6df1\u5c64\u905e\u8ff4\u985e\u795e\u7d93\u7db2\u8def\u96fb\u8996 \u56de\u8072\u6d88\u9664\u7cfb\u7d71\uff0c\u5728\u4ee5\u4e0b\u7ae0\u7bc0\u5c07\u6703\u518d\u8a73\u7d30\u4ecb\u7d39\u5176\u67b6\u69cb\u8207\u8a13\u7df4\u65b9\u6cd5\u3002 \u4e8c\u3001\u76f8\u95dc\u7814\u7a76 \u5e38\u7528\u7684\u56de\u8072\u6d88\u9664\u65b9\u6cd5\u5305\u62ec NLMS\u3001RBF \u5373 MLP\uff0c\u5176\u4e2d\u6700\u5c0f\u5747\u65b9\u6f14\u7b97\u6cd5(LMS)\u7c21\u6613\u5be6\u73fe\u3001 \u8a08\u7b97\u91cf\u5c0f\u53ca\u7a69\u5b9a\u7279\u8272\uff0c\u53d7\u5230\u4e0d\u5c11\u4eba\u9752\u775e\u4e14\u5ee3\u6cdb\u5730\u904b\u7528\uff0c\u540c\u6642\u70ba\u4e86\u89e3\u6c7a\u7cfb\u7d71\u6536\u6582\u7de9\u6162\u7684\u7f3a \u9ede\uff0c\u5728\u4f7f\u7528\u4e0a\u5c07\u8f38\u5165\u8a0a\u865f\u7684\u80fd\u91cf\u9032\u884c\u6b63\u898f\u5283\u8655\u7406\uff0c\u4e5f\u5c31\u662f NLMS \u6f14\u7b97\u6cd5\uff0c\u5176\u63a1\u7528\u53ef\u8b8a \u6b65\u9577\u7684\u65b9\u6cd5\u4f86\u7a69\u5b9a\u6536\u6582\u904e\u7a0b\u3002LMS \u8207 NLMS \u7684\u57fa\u672c\u539f\u7406\u662f\u8a08\u7b97\u8f38\u5165\u4fe1\u865f\u8207\u53c3\u8003\u4fe1\u865f\u7684 \u95dc\u806f\u6027\uff0c\u4e0b\u5716 2 \u70ba NLMS \u81ea\u9069\u61c9\u6ffe\u6ce2\u5668\u67b6\u69cb\u5716\u3002 \u5716 2 NLMS \u81ea\u9069\u61c9\u6ffe\u6ce2\u5668\u67b6\u69cb\u5716 \u5716 3 RBF \u795e\u7d93\u7db2\u8def\u67b6\u69cb\u5716 RBF \u7db2\u8def\u67b6\u69cb\u53ea\u7531\u4e00\u5c64\u7684\u96b1\u85cf\u5c64\u3001\u8f38\u5165\u5c64\u53ca\u8f38\u51fa\u5c64\u5171\u4e09\u5c64\u6240\u69cb\u6210\uff0c\u5176\u6838\u5fc3\u70ba\u9ad8\u65af\u6838\u51fd \u6578\u6240\u7d44\u7e54\uff0c\u8a13\u7df4\u904e\u7a0b\u53ef\u770b\u4f5c\u5728\u9ad8\u7dad\u7a7a\u9593\u4e2d\u627e\u5c0b\u6700\u4f73\u903c\u8fd1\u53c3\u8003\u6578\u64da\u7684\u89e3\uff0c\u4e3b\u8981\u7531\u7b2c\u4e00\u5c64\u8f38 \u5165\u5c64\u7684\u611f\u77e5\u55ae\u5143\u5c07\u795e\u7d93\u7db2\u8def\u8207\u5916\u754c\u76f8\u806f\u63a5\u6536\u8a0a\u606f\uff0c\u4e26\u76f4\u63a5\u50b3\u905e\u5230\u96b1\u85cf\u5c64\uff1b\u63a5\u8457\u7b2c\u4e8c\u5c64\u96b1 \u85cf\u5c64\u5247\u662f\u85c9\u8457\u5c0d\u8f38\u5165\u5411\u91cf\u7a7a\u9593\u5230\u96b1\u85cf\u7a7a\u9593\u4e4b\u9593\u7684\u975e\u7dda\u6027\u6620\u5c04\u8b8a\u63db\uff0c\u800c\u7b2c\u4e09\u5c64\u8f38\u51fa\u5c64\u7d93\u7531 \u7dda\u6027\u7d44\u5408\u52a0\u6b0a\u8b8a\u6210\u8f38\u51fa\u3002 \u800c\u591a\u5c64\u611f\u77e5\u6a5f\u985e\u795e\u7d93\u7db2\u8def\u542b\u6709\u5169\u5c64\u4ee5\u4e0a\u7684\u96b1\u85cf\u5c64\uff0c\u5176\u4e2d\u7528\u5230\u7684\u6fc0\u6d3b\u51fd\u6578\u70ba Sigmoid \u6fc0\u6d3b \u51fd\u6578\uff0c\u6838\u5fc3\u6f14\u7b97\u6cd5\u70ba\u53cd\u5411\u50b3\u64ad\u6f14\u7b97\u6cd5\uff0c\u7528\u65bc\u56de\u8072\u6d88\u9664\u7684\u7db2\u8def\u67b6\u69cb\u5982\u4e0b\u5716 4 \u6240\u793a: \u56db\u3001\u5be6\u9a57\u7d50\u679c 1. \u5be6\u9a57\u4e00\uff0c\u7d14\u96fb\u8996\u7bc0\u76ee\u8072\u6d88\u9664\u5be6\u9a57: \u5716 12 \u5be6\u9a57\u4e09 \u524d\u4e94\u79d2\u55ae\u7d14\u96fb\u8996\u7bc0\u76ee\u8072\u9810\u8a13\u7df4\uff0c\u5f8c\u4e94\u79d2\u6df7\u548c\u4eba\u8072\u96fb\u8996\u56de\u8072\u6d88\u9664\u6a21\u5f0f\uff0c\u5c0d \u4f86\u964d\u4f4e\u4eba\u8072\u6ea2\u5165\u9069\u61c9\u6027\u6ffe\u6ce2\u5668\u7684\u5f71\u97ff\uff0c\u7d50\u679c\u986f\u793a\u4e86\u6b64\u65b9\u6cd5\u5c0d\u65bc\u975e\u7dda\u6027\u6ffe\u6ce2\u6f14\u7b97\u6cd5\u6709\u660e\u986f \u5716 4 \u4e09\u5c64\u591a\u5c64\u611f\u77e5\u6a5f\u7db2\u8def\u67b6\u69cb\u5716 \u4e3b\u8981\u7db2\u8def\u67b6\u69cb\u7531\u8f38\u5165\u5c64\u3001\u591a\u5c64\u96b1\u85cf\u5c64\u53ca\u8f38\u51fa\u5c64\u7d44\u5408\u800c\u6210\uff0c\u7db2\u8def\u904b\u4f5c\u70ba\u8f38\u5165\u5c64\u628a\u8cc7\u8a0a\u50b3\u9001 \u5230\u96b1\u85cf\u5c64\uff0c\u518d\u7531\u6700\u5f8c\u7684\u96b1\u85cf\u5c64\u50b3\u8f38\u5230\u8f38\u51fa\u5c64\uff0c\u52a0\u6b0a\u7e3d\u548c\u51fa\u6574\u500b\u7db2\u8def\u7684\u8f38\u51fa\u503c\uff0c\u6b64\u6642\u70ba\u524d \u994b\u7db2\u8def\uff0c\u800c\u4e14\u6b0a\u91cd\u503c\u53ca\u504f\u58d3\u503c\u662f\u56fa\u5b9a\u503c\uff0c\u5176\u4e2d\u5c64\u8207\u5c64\u4e4b\u9593\u6e9d\u901a\u662f\u795e\u7d93\u5143\u900f\u904e\u6b0a\u91cd\u53ca\u504f\u58d3 \u5716 6 \u6df1\u5c64\u905e\u8ff4\u7db2\u8def\u67b6\u69cb\u5716 (\u4e8c)\u591a\u8072\u9053\u6df1\u5c64\u905e\u8ff4\u985e\u795e\u7d93\u7db2\u8def\u56de\u97f3\u6d88\u9664 \u70ba\u4e86\u63d0\u9ad8\u6536\u97f3\u54c1\u8cea\u4ee5\u53ca\u964d\u4f4e\u6536\u97f3\u89d2\u4f4d\u7684\u5f71\u97ff\uff0c\u6211\u5011\u6539\u9032\u5716 5\uff0c\u52a0\u4e0a\u591a\u96bb\u9ea5\u514b\u98a8\u653e\u5728\u4e0d\u540c \u8868 1 \u8a9e\u6599\u683c\u5f0f\u8a2d\u5b9a\u6574\u7406 \u6f14\u7b97\u6cd5\u5c0d\u96fb\u8996\u7bc0\u76ee\u8072\u6bcf\u4e00\u985e\u7684\u97f3\u6a94\u505a\u56de\u5347\u6d88\u9664\uff0c\u4e26\u505a\u56de\u8072\u8fd4\u56de\u640d\u5931\u5f37\u5316\u6027\u80fd\u6bd4\u8f03\u8a55\u4f30\uff0c \u7528\u4f86\u4e86\u89e3\u6f14\u7b97\u6cd5\u5728\u4e0d\u540c\u985e\u5225\u7684\u96fb\u8996\u7bc0\u76ee\u8072\u8868\u73fe\u5982\u4f55\uff0c\u7136\u5f8c\u770b\u56db\u985e\u6027\u80fd\u6bd4\u8f03\u8a55\u4f30\u505a\u7e3d\u5e73\u5747 \u89c0\u5bdf\u5176\u5c0d\u56db\u985e\u96fb\u8996\u7bc0\u76ee\u8072\u6574\u9ad4\u8868\u73fe\u3002 \u5982\u4e0b\u5716 10 \u6240\u793a\u3002 \u6240\u6709\u4e0d\u540c\u6e2c\u8a66\u4f7f\u7528\u8005\u7684\u56de\u8072\u6d88\u9664 ERLE \u8207 MSE \u7d50\u679c\u5716 (B)\u591a\u901a\u9053\u96fb\u8996\u56de\u8072\u6d88\u9664\u5be6\u9a57 \u7684\u63d0\u5347\u3002\u800c\u5728(B)\u591a\u901a\u9053\u5be6\u9a57\u4e2d\u52a0\u5165\u591a\u901a\u9053\u6df1\u5c64\u905e\u8ff4\u985e\u795e\u7d93\u7db2\u8def\uff0c\u56e0\u5176\u80fd\u9304\u53d6\u4e0d\u540c\u65b9\u4f4d \u7684\u8072\u97f3\u8cc7\u8a0a\uff0c\u6bd4\u55ae\u901a\u9053\u7cfb\u7d71\u5f97\u5230\u66f4\u591a\u7684\u8cc7\u6599\uff0c\u4e5f\u5c31\u80fd\u66f4\u6709\u6548\u7684\u5b78\u7fd2\u96fb\u8996\u7bc0\u76ee\u56de\u8072\u8def\u5f91\uff0c \u7531\u5be6\u9a57\u7d50\u679c\u53ef\u4ee5\u77e5\u9053\uff0c\u591a\u901a\u9053\u6df1\u5c64\u905e\u8ff4\u985e\u795e\u7d93\u7db2\u8def\u56de\u8072\u6d88\u9664\u7cfb\u7d71\u66f4\u597d\u65bc\u55ae\u8072\u9053\u6df1\u5c64\u905e\u8ff4 \u672c\u8ad6\u6587\u7684\u5be6\u9a57\u8a9e\u6599\u5305\u542b 8 \u6a23\u9ede\uff0cRNN \u6f14\u7b97\u6cd5\u8a2d\u5b9a\u70ba\u5169\u5c64\u96b1\u85cf\u5c64\uff0c\u6bcf\u5c64\u795e\u7d93\u5143\u70ba 100\u3002\u5728\u4ee5\u4e0b\u5be6\u9a57\u4e2d\u5148\u6e2c\u8a66(A)\u5168 1. \u5be6\u9a57\u4e00\uff0c\u55ae\u901a\u9053 RNN \u53ca\u591a\u901a\u9053 RNN \u7684\u56de\u8072\u6d88\u9664\u5be6\u9a57 \u985e\u795e\u7d93\u7db2\u8def\u56de\u8072\u6d88\u9664\u7cfb\u7d71\u3002 \u90e8\u65b9\u6cd5\u5728\u55ae\u901a\u9053\u56de\u8072\u6d88\u9664\u7684\u6548\u679c\uff0c\u53d6\u5176\u4e2d\u6700\u597d\u7684\u65b9\u6cd5\u5728(B)\u6e2c\u8a66\u591a\u901a\u9053\u7684\u60c5\u5f62\u3002 \u7d50\u679c\u5982\u4e0b\u5716 13 \u70ba\u524d\u4e94\u79d2\u55ae\u7d14\u80cc\u96fb\u8996\u7bc0\u76ee\u56de\u8072\uff0c\u5f8c\u4e94\u79d2\u6df7\u6709\u4e0d\u540c\u7684\u6e2c\u8a66\u4f7f\u7528\u8005\u7684\u8072\u97f3\uff0c \u4f4d\u7f6e\uff0c\u69cb\u6210\u9ea5\u514b\u98a8\u9663\u5217\uff0c\u5efa\u7acb\u5982\u5716 7 \u7684\u591a\u901a\u9053\u6df1\u5c64\u905e\u8ff4\u985e\u795e\u7d93\u7db2\u8def\u4e4b\u96fb\u8996\u56de\u8072\u6d88\u9664\u7cfb\u7d71 \u503c\u76f8\u4e92\u7684\u9023\u7d50\uff0c\u5f97\u5230\u6574\u500b\u7db2\u8def\u7684\u8f38\u51fa\u503c\u5f8c\uff0c\u5c0d\u6bd4\u76ee\u6a19\u671f\u671b\u7684\u8f38\u51fa\u503c\u7b97\u51fa\u8aa4\u5dee\u503c\uff0c\u900f\u904e\u53cd \u5411\u50b3\u64ad\u6f14\u7b97\u6cd5\u628a\u8aa4\u5dee\u5012\u50b3\u9001\u56de\u795e\u7d93\u7db2\u8def\u4e2d\uff0c\u53bb\u505a\u6b0a\u91cd\u53ca\u504f\u58d3\u503c\u8abf\u6574\uff0c\u6b64\u6642\u70ba\u5012\u50b3\u905e\u7db2 \u8def\uff0c\u5f97\u5230\u6574\u500b\u7db2\u8def\u8f38\u51fa\u5c64\u7684\u503c\u5f8c\uff0c\u8207\u76ee\u6a19\u671f\u671b\u7684\u8f38\u51fa\u503c\u76f8\u6e1b\u7372\u5f97\u8aa4\u5dee\u503c\uff0c\u518d\u5229\u7528\u5176\u8aa4\u5dee \u503c\u6240\u4ee5\u5b9a\u7fa9\u51fa\u4f86\u7684\u6210\u672c\u51fd\u6578\uff0c\u900f\u904e\u53cd\u5411\u50b3\u64ad\u6f14\u7b97\u6cd5\u4f86\u5c0d\u5176\u7db2\u8def\u6b0a\u91cd\u53ca\u504f\u58d3\u503c\u4f5c\u66f4\u65b0\uff0c\u7136 \u5f8c\u518d\u4ee3\u56de\u7db2\u8def\u6c42\u65b0\u8aa4\u5dee\u503c\uff0c\u4f7f\u8aa4\u5dee\u5e73\u65b9\u8da8\u8fd1\u6975\u5c0f\u503c\uff0c\u8b93\u7db2\u8def\u8f38\u51fa\u903c\u8fd1\u65bc\u76ee\u6a19\u671f\u671b\u503c\u3002 \u6700\u5f8c\uff0c\u7531\u4ee5\u4e0a\u8a0e\u8ad6\u53ef\u77e5\uff0cNLMS\uff0cRBF \u8207 MLP \u80fd\u770b\u5230\u7684\u6b77\u53f2\u8a0a\u865f\u53d7\u5230\u5b83\u7684\u6ffe\u6ce2\u5668\u9577\u5ea6\uff0c \u6216\u8f38\u5165\u795e\u7d93\u5143\u6578\u76ee\u9650\u5236\uff0c\u901a\u5e38\u662f\u6709\u9650\u7684\u4e14\u7121\u6cd5\u592a\u9577\u3002 \u4e09\u3001\u57fa\u65bc\u6df1\u5c64\u905e\u8ff4\u985e\u795e\u7d93\u7db2\u8def\u4e4b\u56de\u97f3\u6d88\u9664\u7cfb\u7d71 \u56e0\u70ba\u96fb\u8996\u7bc0\u76ee\u8072\u6703\u5728\u5ba4\u5167\u53cd\u5c04\u9020\u6210\u6b98\u97ff\uff0c\u800c\u4e14\u6b98\u97ff\u6642\u9593\u9577\u5ea6\u5e38\u5e38\u5927\u65bc 0.5 \u79d2\uff0c\u800c\u50b3\u7d71\u6d88 \u9664\u56de\u8072\u7cfb\u7d71\u90fd\u53d7\u9650\u65bc\u8f38\u5165\u9577\u5ea6\u56fa\u5b9a\uff0c\u5982\u679c\u8f38\u5165\u8a2d\u5b9a\u592a\u9577\uff0c\u6703\u5c0e\u81f4\u904b\u7b97\u91cf\u904e\u5927\uff0c\u6536\u6582\u592a\u6162\uff0c \u7121\u6cd5\u6709\u6548\u589e\u5f37\u4f7f\u7528\u8005\u8a9e\u97f3\u8a0a\u606f\uff0c\u6545\u6211\u5011\u6539\u7528\u6df1\u5c64\u905e\u8ff4\u985e\u795e\u7d93\u7db2\u8def[8]\uff0c\u56e0\u5176\u53ef\u4ee5\u7d93\u7531\u56de \u503c\uff0c\u8b93\u7db2\u8def\u8f38\u51fa\u903c\u8fd1\u65bc\u76ee\u6a19\u671f\u671b\u503c\u3002\u4e0b\u5716 6 \u70ba\u672c\u8ad6\u6587\u4e2d\u4f7f\u7528\u7684\u6df1\u5c64 RNN \u904b\u4f5c\u69cb\u9020\u3002 \u5716 7 \u591a\u901a\u9053\u905e\u8ff4\u985e\u795e\u7d93\u7db2\u8def\u96fb\u8996\u56de\u8072\u6d88\u9664\u7cfb\u7d71\u67b6\u69cb \u8072\u6d88\u9664\u7cfb\u7d71\u6d88\u9664\u5f71\u7247\u7bc0\u76ee\u8072\u5f8c\u5f97\u5230\u7684 TCC300 \u6e2c\u8a66\u4eba\u8072\u3002\u85c9\u7531\u539f\u59cb\u7684\u8a0a\u865f\u8072 m(n)\u8207\u7d93 \u6388\u7dda\u8def\u770b\u5230\u5f88\u4e45\u4ee5\u524d\u7684\u8cc7\u6599\u4ee5\u5b78\u7fd2\u9577\u6642\u9593\u7684\u56de\u8072\u8def\u5f91\uff0c\u4e26\u731c\u51fa\u4e0b\u4e00\u6642\u9593\u9ede\u7684\u8072\u97f3\uff0c\u800c\u55ae \u5716 5 \u55ae\u8072\u9053\u6df1\u5c64\u905e\u8ff4\u985e\u795e\u7d93\u7db2\u8def\u7cfb\u7d71\u67b6\u69cb \u6e2c\u8a66\u55ae\u901a\u9053 RNN \u53ca\u591a\u901a\u9053 RNN \u6f14\u7b97\u6cd5\u5e73\u5747\u6548\u80fd\u4ee5\u53ca\u516b\u500b\u4eba\u7684\u7e3d\u5e73\u5747\u3002\u7531\u5716 13 \u5f97\u77e5\uff0c \u67b6\u69cb\uff0c\u5176\u904b\u4f5c\u539f\u7406\u70ba\u904b\u7528 Kinect for Xbox one \u9ea5\u514b\u98a8 4 \u901a\u9053\u6536\u96c6\u5230\u7684\u8a9e\u97f3 mi(n)\uff0c\u518d\u900f (A) \u5728\u55ae\u901a\u9053\u56de\u97f3\u6d88\u9664: \u4e94\u3001\u7d50\u8ad6 \u5716 11 \u5be6\u9a57\u4e8c \u6240\u6709\u4eba\u8072\u6df7\u96fb\u8996\u7bc0\u76ee\u8072\u4ee5 NLMS\uff0cRBF\uff0cMLP \u8207 RNN \u56de\u8072\u6d88\u9664\u7684 ERLE \u591a\u901a\u9053\u6df1\u5c64\u905e\u8ff4\u985e\u795e\u7d93\u7db2\u8def\u56de\u8072\u6d88\u9664\u7cfb\u7d71\u76f8\u7576\u597d\uff0c\u4e14\u66f4\u512a\u65bc\u55ae\u8072\u9053\u6df1\u5c64\u905e\u8ff4\u985e\u795e\u7d93\u7db2\u8def \u904e\u96fb\u8996\u7dda\u8f38\u51fa\u53d6\u5f97\u96fb\u8996\u7bc0\u76ee\u8072 x(n)\uff0c\u5206\u5225\u53bb\u505a\u56db\u6b21\u6df1\u5c64\u905e\u8ff4\u5f0f\u795e\u7d93\u7db2\u8def\uff0c\u5c07\u56de\u8072\u6d88\u9664\u7cfb \u5c07\u9032\u884c\u4e09\u500b\u5be6\u9a57\uff0c\u5be6\u9a57\u4e00\u55ae\u7d14\u5148\u8003\u616e\u53ea\u6709\u96fb\u8996\u7bc0\u76ee\u97f3\uff0c\u800c\u5be6\u9a57\u4e8c\u52a0\u5165\u4eba\u8072\u6df7\u5408\u96fb\u8996\u7bc0\u76ee (\u4e8c) \u8a9e\u6599\u9304\u97f3\u5be6\u9a57\u60c5\u5883\u5047\u8a2d \u5716 9 \u5be6\u969b\u96fb\u8996\u5587\u53ed\u3001\u6536\u97f3\u9ea5\u514b\u98a8\u8207\u5be6\u969b\u6a21\u64ec\u8aaa\u8a71\u8005\u4e4b\u97f3\u6e90\u5587\u53ed\u64fa\u8a2d \u8207 MSE \u7d50\u679c\u5716 \u5728\u672c\u5be6\u9a57\u4e2d\uff0c\u5229\u7528\u4e86\u76e3\u807d\u5f0f\u5587\u53ed\u53ca Kinect \u7b49\u5668\u6750\uff0c\u5be6\u969b\u9304\u97f3\u6a21\u64ec\u667a\u6167\u578b\u96fb\u8996\u64cd\u4f5c\u60c5\u5883 \u56de\u8072\u6d88\u9664\u7cfb\u7d71\u3002 \u7d71\u7684\u8aa4\u5dee\u8a0a\u865f\u52a0\u6b0a\u5e73\u5747\u4ee5\u52a0\u5f37\u8a9e\u97f3\u8a0a\u865f\u3002 \u8072\u3002\u7531\u65bc\u5be6\u9a57\u4e8c\u6709\u53ef\u80fd\u6703\u5b78\u7fd2\u5230\u4eba\u8072\uff0c\u9664\u4e86\u6d88\u9664\u96fb\u8996\u7bc0\u76ee\u56de\u8072\u5916\uff0c\u53ef\u80fd\u4e5f\u628a\u4eba\u8072\u7d66\u6ffe\u6389 \u4e0b\u7684\u56de\u8072\uff0c\u4f5c\u70ba\u96fb\u8996\u7bc0\u76ee\u56de\u8072\u6d88\u9664\u5be6\u9a57\u7684\u8a9e\u6599\uff0c\u63a5\u8457\u5206\u5225\u5148\u5f8c\u5c0e\u5165\u4e86\u7dda\u6027\u6ffe\u6ce2\u7b97 NLMS \u70ba\u4e86\u80fd\u5920\u6a21\u64ec\u771f\u5be6\u7684\u8072\u5b78\u56de\u8072\u6d88\u9664\u7cfb\u7d71\u60c5\u5f62\uff0c\u6211\u5011\u5728\u4e00\u500b\u985e\u4f3c\u5ba2\u5ef3\u7684\u623f\u9593\u6a21\u64ec\u9060\u8ddd\u96e2\u6536 \u5176\u4e2d\u672c\u8ad6\u6587\u63d0\u51fa\u6df1\u5c64\u905e\u8ff4\u985e\u795e\u7d93\u7db2\u8def(RNN)\u67b6\u69cb[6]\u9664\u4e86\u8207\u4e00\u822c\u985e\u795e\u7d93\u7db2\u8def\u4e00\u6a23\u6709\u8f38\u5165 \u4e86\uff0c\u5c0e\u81f4\u4eba\u8072\u8a0a\u865f\u8b8a\u5f62\uff0c\u56e0\u6b64\u518d\u52a0\u5165\u5be6\u9a57\u4e09\uff0c\u5be6\u9a57\u4e09\u70ba\u524d\u4e94\u79d2\u55ae\u7d14\u96fb\u8996\u7bc0\u76ee\u8072\u9810\u8a13\u7df4\u5f8c \u97f3\u60c5\u5883\uff0c\u9996\u5148\uff0c\u628a Kinect for Xbox one \u5145\u7576\u63a5\u6536\u7aef\u9ea5\u514b\u98a8\uff0c\u4e26\u5728 Kinect for Xbox one \u4f4d 3. \u5be6\u9a57\u4e09\uff0c\u524d 5 \u79d2\u7d14\u96fb\u8996\u7bc0\u76ee\u8072\u9810\u8a13\u7df4\u52a0\u5f8c\u4e94\u79d2\u6df7\u4eba\u8072\u8207\u96fb\u8996\u7bc0\u76ee\u56de\u8072\u6d88\u9664\u5be6\u9a57: \u4ee5\u53ca\u975e\u7dda\u6027\u6ffe\u6ce2\u6f14\u7b97\u6cd5\u5982 RBF\u3001MLP\u3001RNN \u7b49\uff0c\u4f5c\u96fb\u8996\u7bc0\u76ee\u56de\u8072\u6d88\u9664\u5be6\u9a57\u3002 (\u4e09) \u56de\u8072\u6d88\u9664\u8a55\u4f30 \u5c64\u3001\u96b1\u85cf\u5c64\u4ee5\u53ca\u8f38\u51fa\u5c64\u5916\uff0c\u9084\u6703\u591a\u51fa\u5169\u500b\u6216\u5169\u500b\u4ee5\u4e0a\u7684\u905e\u8ff4\u5c64\u7576\u56de\u6388\u529f\u80fd\uff0c\u5176\u904b\u4f5c\u539f\u7406 \u4e94\u79d2\u4eba\u8072\u6df7\u96fb\u8996\u7bc0\u76ee\u8072\u56de\u8072\u6d88\u9664\u6a21\u5f0f\uff0c\u5229\u7528\u524d\u4e94\u79d2\u53ea\u6709\u96fb\u8996\u7bc0\u76ee\u8072\uff0c\u53ef\u4ee5\u9810\u5148\u5b78\u7fd2\u74b0\u5883 \u7f6e\u5de6\u53f3\u5169\u65c1\u5e73\u884c\u653e\u4e0a\u5169\u9846\u4e3b\u52d5\u5f0f\u76e3\u807d\u5587\u53ed\uff0c\u7576\u6210\u96fb\u8996\u56de\u8072\u80cc\u666f\u97f3\u7684\u4f86\u6e90\uff0c\u4e14\u5728 Kinect for \u7d50\u679c\u5982\u4e0b\u5716 12 \u70ba\u524d\u4e94\u79d2\u55ae\u7d14\u80cc\u96fb\u8996\u7bc0\u76ee\u56de\u8072\uff0c\u5f8c\u4e94\u79d2\u6df7\u6709\u4e0d\u540c\u7684\u6e2c\u8a66\u4f7f\u7528\u8005\u7684\u8072\u97f3\uff0c \u56de\u8072\u6d88\u9664\u6210\u6548\u9664\u4e86\u4e3b\u89c0\u7684\u7531\u8033\u6735\u807d\u53d6\u8072\u97f3\u5916\uff0c\u9084\u53ef\u4ee5\u7528\u5e73\u5747\u8aa4\u5dee\u503c(Mean Squared \u4e3b\u8981\u900f\u904e\u56de\u6388\u7dda\u8def\uff0c\u628a\u4e0a\u500b\u6642\u9593\u9ede\u7684\u96b1\u85cf\u5c64\u795e\u7d93\u5143\u8f38\u51fa\u503c\u8a18\u9304\u4e0b\u4f86\uff0c\u4e26\u91cd\u65b0\u5c0e\u5165\u5230\u96b1\u85cf \u97ff\u61c9\uff0c\u9019\u6a23\u4e0d\u6703\u5b78\u5230\u4f7f\u7528\u8005\u7684\u4eba\u8072\uff0c\u800c\u5f8c\u4e94\u79d2\u5247\u628a\u5b78\u7fd2\u7387\u8abf\u4f4e\uff0c\u964d\u4f4e\u6ffe\u9664\u4eba\u8072\u7684\u53ef\u80fd\u6027\uff0c Xbox one \u7684\u6b63\u524d\u65b9\u8ddd\u96e2 2m \u8655\u4e5f\u64fa\u4e0a\u4e3b\u52d5\u5f0f\u76e3\u807d\u5587\u53ed\u64ad\u653e\u51fa\u4eba\u8072\uff0c\u6a21\u64ec\u4f7f\u7528\u8005\u6b63\u5728\u8b1b Error\uff0cMSE)\u8207\u56de\u8072\u8fd4\u56de\u640d\u5931\u5f37\u5316((Echo Return Loss Enhancement\uff0cERLE)[8]\u6578\u503c\u5316\u5728\u6642 \u6e2c\u8a66\u6f14\u7b97\u6cd5\u5e73\u5747\u6548\u80fd\u4ee5\u53ca\u516b\u500b\u4eba\u7684\u7e3d\u5e73\u5747\u3002\u7531\u5716 12 \u660e\u986f\u5f97\u77e5\u6b64\u6642 NLMS \u8868\u73fe\u6700\u5dee\uff0c\u55ae \u7d93\u7531\u5be6\u9a57\u7d50\u679c\u53ef\u77e5\uff0c\u5728\u5927\u90e8\u5206\u60c5\u6cc1\u4e0b\uff0cRNN \u7684\u8868\u73fe\u76f8\u7576\u7a69\u5b9a\uff0c\u5c24\u5176\u662f\u591a\u901a\u9053\u6df1\u5c64\u905e\u56de \u5716 10 \u5be6\u9a57\u4e00 \u6240\u6709\u96fb\u8996\u7bc0\u76ee\u985e\u578b\u4ee5 NLMS\uff0cRBF\uff0cMLP \u8207 RNN \u56de\u8072\u6d88\u9664\u7684 ERLE \u8207 \u5c64\u795e\u7d93\u5143\u8f38\u5165\u7aef\uff1b\u8207\u8f38\u5165\u5c64\u8f38\u5165\u503c\u6574\u5408\u5728\u4e00\u8d77\uff0c\u7576\u4e0b\u4e00\u500b\u6642\u9593\u9ede\u7684\u96b1\u85cf\u5c64\u8f38\u5165\uff0c\u4ee5\u6b64\u8b93 \u9054\u5230\u6d88\u9664\u96fb\u8996\u56de\u8072\u4e26\u589e\u5f37\u4f7f\u7528\u8005\u7684\u8a9e\u97f3\u8a0a\u865f\u3002\u5728\u6240\u6709\u5be6\u9a57\u4e2d\uff0c\u6211\u5011\u900f\u904e\u56de\u8072\u8fd4\u56de\u640d\u5931\u5f37 \u8a71\uff0c\u5982\u6b64\u4e00\u4f86\uff1b\u53ef\u4ee5\u60f3\u50cf\u51fa\uff0c\u7576\u64ad\u653e\u51fa\u4eba\u8072\u6642\uff0c\u5f71\u7247\u7bc0\u76ee\u8072\u4e5f\u540c\u6642\u6df7\u9032\u758a\u52a0\u5176\u4e2d\uff0c\u4e00\u8d77 \u57df\u4e0a\u7684\u5dee\u7570\uff0c\u4f5c\u70ba\u56de\u8072\u6d88\u9664\u5f8c\u7684\u8a55\u4f30\u5176\u65b9\u7a0b\u5f0f\u5982\u4e0b\u6240\u793a: \u8072\u9053\u6df1\u5c64\u905e\u8ff4\u985e\u795e\u7d93\u7db2\u8def\u56de\u8072\u6d88\u9664\u7cfb\u7d71\u5728\u7e3d\u5e73\u5747\u8981\u597d\u65bc RBF \u53ca MLP\u3002 \u985e\u795e\u7d93\u7db2\u8def\u56de\u8072\u6d88\u9664\u7684\u6548\u679c\u53ef\u4ee5\u9054\u5230\u6700\u4f73\u3002\u672c\u8ad6\u6587\u5be6\u9a57\u7d50\u679c\uff0c\u53ef\u4ee5\u70ba\u65e5\u5f8c\u56de\u8072\u6d88\u9664\u7814\u7a76 MSE \u7d50\u679c\u5716 \u795e\u7d93\u5143\u4e0b\u500b\u6642\u9593\u8f38\u5165\u503c\u8207\u904e\u53bb\u8f38\u51fa\u503c\u6709\u95dc\uff0c\u7c21\u55ae\u4f86\u8aaa\u8b93\u6574\u500b\u795e\u7d93\u7db2\u8def\u662f\u6709\u8a18\u61b6\u6027\u7684\uff0c\u4e5f \u5316\u4f5c\u70ba\u56de\u8072\u6d88\u9664\u5f8c\u7684\u8a55\u4f30\uff0c\u4f86\u6bd4\u8f03\u56db\u7a2e\u6f14\u7b97\u6cd5\u7684\u597d\u58de\uff0c\u4e26\u9078\u51fa\u6700\u597d\u7684\u6f14\u7b97\u6cd5\u3002 \u63d0\u4f9b\u53c3\u8003\uff0c\u76f8\u4fe1\u672a\u4f86\u4ecd\u6709\u8a31\u591a\u53ef\u4ee5\u6539\u9032\u7684\u7a7a\u9593\uff0c\u4f8b\u5982\uff1a\u6df1\u5c64\u985e\u795e\u7d93\u7db2\u8def\u589e\u52a0\u5c64\u6578\u4ee5\u53ca\u6bcf \u88ab\u9ea5\u514b\u98a8\u9663\u5217\u6240\u63a5\u6536\u4e26\u9304\u97f3\u8d77\u4f86\uff0c\u9019\u6a23\u53ef\u4ee5\u7528\u4f86\u7576\u6211\u5011\u8072\u5b78\u56de\u8072\u6d88\u9664\u7cfb\u7d71\u7684\u8a9e\u6599\u4e86\u3002\u5be6 \u53ef\u4ee5\u8aaa\u662f\u628a\u8f38\u5165\u5c64\u7d66\u653e\u5927\u4e86\uff0c\u5176\u6838\u5fc3\u53c3\u6578\u66f4\u65b0\u6f14\u7b97\u6cd5\u9084\u662f\u53cd\u5411\u50b3\u64ad\u6f14\u7b97\u6cd5\uff0c\u4f46\u4e0d\u4e00\u6a23\u7684 \u969b\u4e0a\u6211\u5011\u5171\u9304\u88fd\u8aaa\u8a71\u8005\u5728 90\uff0c60 \u8207 30 \u5ea6\u89d2\u4f4d\u7f6e\u7684\u4eba\u8072\uff0c\u4f46\u76ee\u524d\u672c\u8ad6\u6587\u5be6\u9a57\u56de\u8072\u6d88\u9664\u53ea ERLE= (1) \u4e00\u5c64\u7db2\u8def\u795e\u7d93\u5143\u6578\u7684\u8abf\u6574\u7b49\u7b49\uff0c\u9084\u53ef\u4ee5\u591a\u6a21\u64ec\u4e00\u4e9b\u60c5\u6cc1\uff0c\u4f86\u4e86\u89e3\u5c0d\u65bc\u56de\u8072\u6d88\u9664\u7cfb\u7d71\u6709\u4ec0 \u5728\u5be6\u9a57\u4e00\u7d14\u96fb\u8996\u7bc0\u76ee\u8072\u6d88\u9664\u5be6\u9a57\u4e0a\uff0c\u6574\u9ad4\u4f86\u8aaa\u97f3\u6a02\u985e\u6700\u5bb9\u6613\u8655\u7406\uff0c\u65b0\u805e\u985e\u6700\u96e3\u8655\u7406\u3002\u6b64 \u5730\u65b9\u5728\u65bc\u9700\u8981\u6839\u64da\u6642\u9593\u5148\u5f8c\u9806\u5e8f\u4f86\u505a\u6b0a\u91cd\u8abf\u6574\uff0c\u6240\u4ee5\u6b0a\u91cd\u8abf\u6574\u6703\u900f\u904e\u4e0d\u540c\u6642\u9593\u9ede\u7684\u96b1\u85cf (B) \u591a\u901a\u9053\u56de\u97f3\u6d88\u9664\u5be6\u9a57\u4e2d: \u6709\u62ff 90 \u5ea6\u89d2\u7684 TCC300 \u6e2c\u8a66\u4eba\u8072\u9304\u97f3\u8a9e\u6599\u4f86\u4f7f\u7528\uff0c\u66ab\u4e0d\u8003\u616e\u5176\u4ed6\u89d2\u5ea6\uff0c\u5176\u4f48\u5c40\u64fa\u8a2d\u5982 \u9ebc\u5f71\u97ff\uff0c\u50cf\u89d2\u5ea6\u56de\u8072\u7684\u5f71\u97ff\u4ee5\u53ca\u591a\u4eba\u4f7f\u7528\u8005\u60c5\u6cc1\u4e0b\u7b49\u7b49\uff0c\u9019\u4e9b\u90fd\u662f\u65e5\u5f8c\u53ef\u4ee5\u52a0\u4ee5\u8003\u616e\u7684 \u5916\uff0cRNN \u5728\u5927\u90e8\u5206\u60c5\u6cc1\u4e0b\u7684\u8868\u73fe\u90fd\u76f8\u7576\u4e0d\u932f\uff0c\u6240\u4ee5\u5176\u7e3d\u5e73\u5747\u5206\u6578\u512a\u65bc\u6240\u6709\u65b9\u6cd5\u3002 \u5c64\u8a0a\u606f\u9032\u884c\uff0c\u5148\u7531\u6700\u5f8c\u6642\u9593\u9ede\u958b\u59cb\u5c0d\u65bc\u6210\u672c\u51fd\u6578\u4f5c\u504f\u5fae\u5206\uff0c\u5f80\u524d\u7b97\u51fa\u5230\u4e00\u958b\u59cb\u6642\u9593\u9ede\u7684 \u504f\u5fae\u5206\u503c\uff0c\u76f4\u5230\u6574\u500b\u6b0a\u91cd\u4f5c\u51fa\u8abf\u6574\u5b8c\u5f8c\uff0c\u518d\u4ee3\u56de\u7db2\u8def\u6c42\u65b0\u8aa4\u5dee\u503c\uff0c\u4f7f\u8aa4\u5dee\u5e73\u65b9\u8da8\u8fd1\u6975\u5c0f \u4f9d\u64da\u4e4b\u524d\u5be6\u9a57\u4e09\u7684\u8a2d\u5b9a\uff0c\u9032\u884c\u591a\u901a\u9053\u56de\u97f3\u6d88\u9664\u5be6\u9a57\uff0c\u4ee5\u5728\u55ae\u901a\u9053\u8868\u73fe\u6700\u597d\u7684 RNN \u4f9d\u5f8c \u9762\u7684\u5be6\u9a57\u7d50\u679c\uff0c\u88fd\u4f5c\u591a\u901a\u9053\u56de\u8072\u6d88\u9664\uff0c\u4e26\u8207\u55ae\u901a\u9053 RNN \u7684\u56de\u8072\u6d88\u9664\u5be6\u9a57\u6bd4\u8f03\u3002 \u7528\u8005\u7684\u56de\u8072\u6d88\u9664 ERLE \u8207 MSE \u7d50\u679c\u5716 \u5176\u4e2d\uff0cm(n)\u70ba TCC300 \u6e2c\u8a66\u4eba\u8072\u6df7\u96fb\u8996\u7bc0\u76ee\u8072\u7684\u9304\u97f3\u8a0a\u865f\u8072\u3002s(n)\u70ba\u8aa4\u5dee\u8a0a\u865f\uff1b\u5373\u7d93\u56de \u4e0b\u5716 8 \u53ca 9\u3002 \u5716 13 \u5be6\u9a57\u4e00 \u6bd4\u8f03\u55ae\u901a\u9053 RNN \u8207\u591a\u901a\u9053 RNN \u96fb\u8996\u56de\u8072\u6d88\u9664\u6a21\u5f0f\uff0c\u5c0d\u6240\u6709\u4e0d\u540c\u6e2c\u8a66\u4f7f \u56e0\u7d20\u3002</td></tr><tr><td>374 377</td></tr></table>", | |
| "text": "The 2016 Conference on Computational Linguistics and Speech Processing ROCLING 2016, pp. 372-386 \uf0d3 The Association for Computational Linguistics and Chinese Language Processing \u500b\u8a9e\u8005 TCC300 \u6e2c\u8a66\u4eba\u8072\u4ee5\u53ca 4 \u985e 40 \u500b\u96fb\u8996\u7bc0\u76ee\u8072\uff0c\u6e2c\u8a66 NLMS\u3001RBF\u3001MLP \u4ee5\u53ca RNN \u7b49\u56db\u7a2e\u9069\u61c9\u6027\u6ffe\u6ce2\u5668\u6f14\u7b97\u6cd5\u3002\u5176\u4e2d\u8f38\u5165\u97f3\u6846\u70ba 2048 \u7684\u53d6", | |
| "type_str": "table", | |
| "html": null | |
| } | |
| } | |
| } | |
| } |