{"id":"https://openalex.org/W4308443027","doi":"https://doi.org/10.1049/sil2.12173","title":"A novel end\u2010to\u2010end deep separation network based on attention mechanism for single channel blind separation in wireless communication","display_name":"A novel end\u2010to\u2010end deep separation network based on attention mechanism for single channel blind separation in wireless communication","publication_year":2022,"publication_date":"2022-11-07","ids":{"openalex":"https://openalex.org/W4308443027","doi":"https://doi.org/10.1049/sil2.12173"},"language":"en","primary_location":{"is_oa":true,"landing_page_url":"https://doi.org/10.1049/sil2.12173","pdf_url":null,"source":{"id":"https://openalex.org/S30889260","display_name":"IET Signal Processing","issn_l":"1751-9675","issn":["1751-9675","1751-9683"],"is_oa":true,"is_in_doaj":true,"is_core":true,"host_organization":"https://openalex.org/P4310311714","host_organization_name":"Institution of Engineering and Technology","host_organization_lineage":["https://openalex.org/P4310311714"],"host_organization_lineage_names":["Institution of Engineering and Technology"],"type":"journal"},"license":"cc-by-nc-nd","license_id":"https://openalex.org/licenses/cc-by-nc-nd","version":"publishedVersion","is_accepted":true,"is_published":true},"type":"article","type_crossref":"journal-article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://doi.org/10.1049/sil2.12173","any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5009211676","display_name":"Hao Ma","orcid":"https://orcid.org/0000-0002-4915-5893"},"institutions":[{"id":"https://openalex.org/I4210163363","display_name":"PLA Army Engineering University","ror":"https://ror.org/05mgp8x93","country_code":"CN","type":"education","lineage":["https://openalex.org/I4210163363"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Hao Ma","raw_affiliation_strings":["School of Communication Engineering Army Engineering University of PLA Nanjing China"],"affiliations":[{"raw_affiliation_string":"School of Communication Engineering Army Engineering University of PLA Nanjing China","institution_ids":["https://openalex.org/I4210163363"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5042329532","display_name":"Xiang Zheng","orcid":"https://orcid.org/0000-0002-2990-2169"},"institutions":[{"id":"https://openalex.org/I4210163363","display_name":"PLA Army Engineering University","ror":"https://ror.org/05mgp8x93","country_code":"CN","type":"education","lineage":["https://openalex.org/I4210163363"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Xiang Zheng","raw_affiliation_strings":["School of Communication Engineering Army Engineering University of PLA Nanjing China"],"affiliations":[{"raw_affiliation_string":"School of Communication Engineering Army Engineering University of PLA Nanjing China","institution_ids":["https://openalex.org/I4210163363"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5106403416","display_name":"Lu Yu","orcid":"https://orcid.org/0000-0002-0550-7754"},"institutions":[{"id":"https://openalex.org/I4210163363","display_name":"PLA Army Engineering University","ror":"https://ror.org/05mgp8x93","country_code":"CN","type":"education","lineage":["https://openalex.org/I4210163363"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Lu Yu","raw_affiliation_strings":["School of Communication Engineering Army Engineering University of PLA Nanjing China"],"affiliations":[{"raw_affiliation_string":"School of Communication Engineering Army Engineering University of PLA Nanjing China","institution_ids":["https://openalex.org/I4210163363"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5011261173","display_name":"Xingyu Zhou","orcid":"https://orcid.org/0000-0003-3439-4450"},"institutions":[{"id":"https://openalex.org/I4210163363","display_name":"PLA Army Engineering University","ror":"https://ror.org/05mgp8x93","country_code":"CN","type":"education","lineage":["https://openalex.org/I4210163363"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Xingyu Zhou","raw_affiliation_strings":["School of Communication Engineering Army Engineering University of PLA Nanjing China"],"affiliations":[{"raw_affiliation_string":"School of Communication Engineering Army Engineering University of PLA Nanjing China","institution_ids":["https://openalex.org/I4210163363"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5100628854","display_name":"Yufan Chen","orcid":null},"institutions":[{"id":"https://openalex.org/I4210163363","display_name":"PLA Army Engineering University","ror":"https://ror.org/05mgp8x93","country_code":"CN","type":"education","lineage":["https://openalex.org/I4210163363"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yufan Chen","raw_affiliation_strings":["School of Communication Engineering Army Engineering University of PLA Nanjing China"],"affiliations":[{"raw_affiliation_string":"School of Communication Engineering Army Engineering University of PLA Nanjing China","institution_ids":["https://openalex.org/I4210163363"]}]}],"institution_assertions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":{"value":2000,"currency":"EUR","value_usd":2200,"provenance":"doaj"},"apc_paid":{"value":2000,"currency":"EUR","value_usd":2200,"provenance":"doaj"},"fwci":0.456,"has_fulltext":false,"cited_by_count":2,"citation_normalized_percentile":{"value":0.389181,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":71,"max":77},"biblio":{"volume":"17","issue":"2","first_page":null,"last_page":null},"is_retracted":false,"is_paratext":false,"primary_topic":{"id":"https://openalex.org/T11447","display_name":"Blind Source Separation and Independent Component Analysis","score":0.9998,"subfield":{"id":"https://openalex.org/subfields/1711","display_name":"Signal Processing"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T11447","display_name":"Blind Source Separation and Independent Component Analysis","score":0.9998,"subfield":{"id":"https://openalex.org/subfields/1711","display_name":"Signal Processing"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10860","display_name":"Speech Enhancement Techniques","score":0.9998,"subfield":{"id":"https://openalex.org/subfields/1711","display_name":"Signal Processing"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T12131","display_name":"Deep Learning for Wireless Signal Classification","score":0.9982,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/end-to-end-principle","display_name":"End-to-end principle","score":0.57694167},{"id":"https://openalex.org/keywords/blind-separation","display_name":"Blind Separation","score":0.55603},{"id":"https://openalex.org/keywords/signal-decomposition","display_name":"Signal Decomposition","score":0.545383}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.75274193},{"id":"https://openalex.org/C183115368","wikidata":"https://www.wikidata.org/wiki/Q856577","display_name":"Weighting","level":2,"score":0.58267844},{"id":"https://openalex.org/C74296488","wikidata":"https://www.wikidata.org/wiki/Q2527392","display_name":"End-to-end principle","level":2,"score":0.57694167},{"id":"https://openalex.org/C120317606","wikidata":"https://www.wikidata.org/wiki/Q17105967","display_name":"Blind signal separation","level":3,"score":0.48636678},{"id":"https://openalex.org/C118505674","wikidata":"https://www.wikidata.org/wiki/Q42586063","display_name":"Encoder","level":2,"score":0.452668},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.43855435},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.41061273},{"id":"https://openalex.org/C127162648","wikidata":"https://www.wikidata.org/wiki/Q16858953","display_name":"Channel (broadcasting)","level":2,"score":0.40488884},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.32613057},{"id":"https://openalex.org/C76155785","wikidata":"https://www.wikidata.org/wiki/Q418","display_name":"Telecommunications","level":1,"score":0.1743168},{"id":"https://openalex.org/C71924100","wikidata":"https://www.wikidata.org/wiki/Q11190","display_name":"Medicine","level":0,"score":0.0},{"id":"https://openalex.org/C126838900","wikidata":"https://www.wikidata.org/wiki/Q77604","display_name":"Radiology","level":1,"score":0.0},{"id":"https://openalex.org/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"is_oa":true,"landing_page_url":"https://doi.org/10.1049/sil2.12173","pdf_url":null,"source":{"id":"https://openalex.org/S30889260","display_name":"IET Signal Processing","issn_l":"1751-9675","issn":["1751-9675","1751-9683"],"is_oa":true,"is_in_doaj":true,"is_core":true,"host_organization":"https://openalex.org/P4310311714","host_organization_name":"Institution of Engineering and Technology","host_organization_lineage":["https://openalex.org/P4310311714"],"host_organization_lineage_names":["Institution of Engineering and Technology"],"type":"journal"},"license":"cc-by-nc-nd","license_id":"https://openalex.org/licenses/cc-by-nc-nd","version":"publishedVersion","is_accepted":true,"is_published":true}],"best_oa_location":{"is_oa":true,"landing_page_url":"https://doi.org/10.1049/sil2.12173","pdf_url":null,"source":{"id":"https://openalex.org/S30889260","display_name":"IET Signal Processing","issn_l":"1751-9675","issn":["1751-9675","1751-9683"],"is_oa":true,"is_in_doaj":true,"is_core":true,"host_organization":"https://openalex.org/P4310311714","host_organization_name":"Institution of Engineering and Technology","host_organization_lineage":["https://openalex.org/P4310311714"],"host_organization_lineage_names":["Institution of Engineering and Technology"],"type":"journal"},"license":"cc-by-nc-nd","license_id":"https://openalex.org/licenses/cc-by-nc-nd","version":"publishedVersion","is_accepted":true,"is_published":true},"sustainable_development_goals":[],"grants":[],"datasets":[],"versions":[],"referenced_works_count":26,"referenced_works":["https://openalex.org/W2023620691","https://openalex.org/W2101416065","https://openalex.org/W2161062000","https://openalex.org/W2194775991","https://openalex.org/W2272847350","https://openalex.org/W2348271723","https://openalex.org/W2475942637","https://openalex.org/W2555261587","https://openalex.org/W2597592385","https://openalex.org/W2734774145","https://openalex.org/W2791236149","https://openalex.org/W2808746047","https://openalex.org/W2884585870","https://openalex.org/W2962715207","https://openalex.org/W2962935966","https://openalex.org/W2963204961","https://openalex.org/W2974817461","https://openalex.org/W2982220924","https://openalex.org/W3008163934","https://openalex.org/W3015199127","https://openalex.org/W3026010377","https://openalex.org/W3029444444","https://openalex.org/W3041277184","https://openalex.org/W3099330747","https://openalex.org/W3112812252","https://openalex.org/W3133514748"],"related_works":["https://openalex.org/W4372260270","https://openalex.org/W4206560911","https://openalex.org/W2787352659","https://openalex.org/W2180954594","https://openalex.org/W2127804977","https://openalex.org/W2108418243","https://openalex.org/W2052835778","https://openalex.org/W2049003611","https://openalex.org/W1970611213","https://openalex.org/W164103134"],"abstract_inverted_index":{"IET":[0],"Signal":[1],"ProcessingVolume":[2],"17,":[3],"Issue":[4],"2":[5,1436,1626,2028,2137,2416,2452,2471,2476,2738,2776,2846,2866,2931,2948,2961,2978,2990,3394,3419,3529,3610,3640,3763,3783,3965,3970,3983,4098],"e12173":[6],"ORIGINAL":[7],"RESEARCHOpen":[8],"Access":[9],"A":[10,1227,2714,3360,3381,3595,3602,3655,3686],"novel":[11,1228,2260],"end-to-end":[12,374,1051],"deep":[13,375,423,803,843,857,1052,1093,1138,1229,1263,1336,1398,2101,2207,4147],"separation":[14,24,376,410,424,519,530,573,632,682,735,758,785,798,816,881,925,954,979,1019,1029,1053,1094,1139,1185,1216,1301,1337,1357,1399,1420,1554,1821,1924,2066,3858,4158],"network":[15,377,425,860,869,911,951,1020,1054,1264,1338,1411,1421,1555,1822,1925,2353,2396,2524],"based":[16,378,1055,1277,4032],"on":[17,379,1056,1278,3133,4033],"attention":[18,381,1058,1201,2264,2654,3255,3357],"mechanism":[19,382,1059,2265,2655],"for":[20,44,86,106,126,146,168,210,230,250,270,360,405,480,571,864,883,1075,1247,1323,1485,1869,3030,3054,3855,4119],"single":[21],"channel":[22,475,740,1207,1317,1513,1785,1793,2833,3104,3163,3180,3313,3338,3441,3573,4196,4218],"blind":[23,529,784,3856],"in":[25,414,507,536,595,789,809,839,922,1007,1224,1345,1386,1392,1422,1452,1511,1583,1768,1813,1916,1953,2056,2105,2288,2302,2529,2539,2582,2586,3067,3071,3866,4072,4097,4130,4194],"wireless":[26,364],"communication":[27,365,540,822,848,884,1250,1425,1448],"Hao":[28,30,152,154],"Ma,":[29,153],"Ma":[31,155],"orcid.org/0000-0002-4915-5893":[32,156],"School":[33,57,71,95,115,135,157,181,195,219,239,259],"of":[34,40,58,64,72,78,96,102,116,122,136,142,158,164,182,188,196,202,220,226,240,246,260,266,301,311,324,335,411,420,434,464,490,520,534,539,556,574,604,661,677,697,703,716,730,748,761,821,842,875,900,924,936,955,963,981,1044,1136,1144,1180,1208,1294,1308,1331,1358,1368,1381,1396,1403,1409,1418,1433,1444,1499,1589,1660,1682,1765,1783,1819,1852,1922,1978,2044,2050,2080,2111,2162,2216,2252,2283,2295,2332,2351,2359,2370,2379,2426,2522,2534,2567,2606,2668,2700,2708,2761,2812,2851,3047,3098,3116,3121,3128,3140,3147,3193,3211,3334,3372,3465,3534,3578,3616,3646,3708,3816,3820,3826,3862,4090,4122,4178,4213],"Communication":[35,59,73,97,117,137,159,183,197,221,241,261],"Engineering,":[36,60,74,98,118,138,160,184,198,222,242,262],"Army":[37,61,75,99,119,139,161,185,199,223,243,263],"Engineering":[38,62,76,100,120,140,162,186,200,224,244,264],"University":[39,63,77,101,121,141,163,187,201,225,245,265],"PLA,":[41,65,79,103,123,143,165,189,203,227,247,267],"Nanjing,":[42,66,104,124,144,166,190,228,248,268],"ChinaSearch":[43,105,125,145,167,229,249,269],"more":[45,87,107,127,147,169,211,231,251,271,4155],"papers":[46,88,108,128,148,170,212,232,252,272],"by":[47,89,109,129,149,171,213,233,253,273,706,958,1033,1048,1086,1846,1911,1961,2219,2348,2572,2580,2615,2771,3265,3344,3508,3590,3670,3740,4026],"this":[48,90,110,130,150,172,214,234,254,274,336,371,2257,2694,4073],"authorXiang":[49,173],"Zheng,":[50,70,174,194],"Corresponding":[51,175],"Author":[52,176],"Xiang":[53,69,177,193],"Zheng":[54,178],"[email":[55,84,179,208],"protected]":[56,180],"China":[67,191],"Correspondence":[68,192],"Nanjing":[80,204],"210000,":[81,205],"China.":[82,206],"Email:":[83,207],"protected]Search":[85,209],"authorLu":[91,215],"Yu,":[92,216],"Lu":[93,217],"Yu":[94,218],"authorXingyu":[111,235],"Zhou,":[112,236],"Xingyu":[113,237],"Zhou":[114,238],"authorYufan":[131,255],"Chen,":[132,256],"Yufan":[133,257],"Chen":[134,258,891],"author":[151,275],"First":[276],"published:":[277],"07":[278],"November":[279],"2022":[280],"https://doi.org/10.1049/sil2.12173AboutSectionsPDF":[281],"ToolsRequest":[282],"permissionExport":[283],"citationAdd":[284],"to":[285,307,330,390,408,456,473,553,564,606,653,664,673,721,726,755,773,793,847,861,887,913,984,998,1115,1127,1165,1173,1183,1197,1203,1242,1298,1377,1536,1556,1594,1838,1863,1890,1967,2014,2064,2073,2091,2099,2203,2211,2230,2239,2271,2277,2315,2339,2462,2482,2493,2498,2505,2628,2685,2755,2795,2798,2823,3094,3161,3205,3215,3229,3252,3324,3458,3462,3483,3516,3564,3666,3725,3732,3804,4016],"favoritesTrack":[286],"citation":[287],"ShareShare":[288],"Give":[289],"accessShare":[290,293],"full":[291],"text":[292],"full-text":[294,309,333],"accessPlease":[295],"review":[296],"our":[297,4107],"Terms":[298,321],"and":[299,303,315,322,341,357,450,461,477,616,625,642,666,699,759,796,818,829,877,903,995,1023,1119,1156,1177,1206,1282,1291,1406,1482,1579,1611,1670,1679,1792,1828,1950,1975,2077,2335,2346,2398,2440,2518,2574,2598,2624,2661,2680,2689,2730,3023,3051,3101,3164,3188,3249,3257,3352,3624,3659,3714,3751,3813,3823,3851,3874,3891,4071,4141,4145,4170,4191,4215],"Conditions":[300,323],"Use":[302],"check":[304],"box":[305],"below":[306,329],"share":[308,331],"version":[310,334],"article.I":[312],"have":[313,648,969,1304,2593,2600,3136],"read":[314],"accept":[316],"the":[317,327,380,387,392,398,406,412,421,427,446,459,465,470,481,488,491,498,504,512,518,537,554,566,572,577,590,596,607,612,617,650,659,675,681,695,701,708,714,717,723,739,746,752,766,778,782,840,872,919,933,953,960,978,1004,1008,1015,1035,1041,1045,1057,1063,1067,1070,1080,1087,1091,1096,1099,1107,1110,1116,1120,1123,1128,1134,1137,1142,1145,1149,1157,1167,1181,1194,1199,1209,1215,1225,1235,1244,1261,1279,1289,1295,1300,1306,1309,1324,1329,1332,1342,1350,1356,1369,1378,1382,1394,1397,1401,1407,1410,1416,1419,1431,1434,1456,1459,1486,1497,1500,1512,1517,1520,1531,1538,1553,1558,1569,1595,1601,1658,1661,1676,1683,1694,1701,1709,1758,1769,1780,1800,1816,1820,1847,1853,1860,1865,1877,1885,1892,1912,1923,1943,1973,1979,1991,2009,2034,2042,2051,2060,2068,2075,2081,2119,2160,2163,2213,2217,2221,2226,2235,2240,2253,2263,2273,2278,2284,2333,2337,2349,2357,2368,2371,2376,2387,2400,2424,2427,2435,2441,2464,2483,2494,2500,2506,2515,2520,2523,2603,2607,2611,2616,2621,2640,2645,2665,2669,2677,2681,2698,2705,2709,2757,2762,2768,2773,2809,2813,2825,2837,2852,3037,3045,3059,3080,3087,3096,3103,3108,3114,3117,3122,3126,3129,3138,3145,3148,3151,3154,3194,3200,3212,3222,3254,3259,3286,3319,3326,3332,3373,3432,3445,3452,3466,3470,3477,3509,3518,3535,3579,3617,3647,3706,3709,3715,3727,3733,3741,3746,3755,3797,3811,3821,3834,3844,3848,3852,3863,3869,3878,4005,4011,4018,4027,4034,4068,4075,4088,4091,4115,4120,4123,4131,4146,4157,4188,4195,4206],"Wiley":[318],"Online":[319],"Library":[320],"UseShareable":[325],"LinkUse":[326],"link":[328],"a":[332,347,401,415,562,687,855,896,905,943,1103,1489,1735,1753,1762,1839,1905,1983,2016,2153,2259,2568,2575,2594,2800,3027,3217,3349,3600,3630,3805,4210],"article":[337],"with":[338,438,511,523,670,751,777,974,1320,1349,1361,1514,1761,1884,2262,3199,3318,3451,3745,4162,4197,4209],"your":[339],"friends":[340],"colleagues.":[342],"Learn":[343],"more.Copy":[344],"URL":[345],"Share":[346],"linkShare":[348],"onFacebookTwitterLinkedInRedditWechat":[349],"Abstract":[350],"The":[351,394,418,441,493,620,868,916,949,966,1028,1131,1312,1334,1366,1375,1441,1586,1830,1849,1919,2047,2108,2122,2191,2206,2281,2291,2305,2342,2405,2509,2532,2542,2561,2635,2648,2671,2790,3039,3075,3168,3191,3269,3312,3337,3356,3539,3559,3585,3672,3697,3721,3736,3757,3896,4022,4036,4082,4126,4181,4199],"traditional":[352,505,578,613,621,783,975,1259,1343],"methods":[353,506,579,629,652,845,1344],"exhibit":[354,630],"unstable":[355,797],"performance":[356,633,702,882,926,980,1030],"high":[358,794],"complexity":[359,795,972],"separating":[361,508,1346,1423],"co-frequency":[362,575,1248,1424,1446,1590],"modulated":[363,1249,1447,1591,4124],"signals":[366,510,522,591,776,823,874,885,902,1082,1251,1348,1360,1449,1505,2242,3850,3854,4161,4183,4190],"under":[367,634,1062,1234,4173],"single-channel":[368,635],"conditions.":[369],"In":[370,800,1066,1090,1141,1964,2256,2693,3150,3832,3860,4106],"study,":[372],"an":[373,992,1050,2195,2317,2630,2653,3353,4216],"is":[383,426,431,533,1031,1060,1133,1188,1232,1239,1275,1371,1384,1430,1450,1477,1528,1549,1573,1704,1835,1856,1882,1901,1952,2012,2054,2126,2152,2193,2266,2269,2286,2310,2344,2365,2431,2491,2652,2711,2753,2793,2821,3026,3172,3204,3227,3342,3456,3481,3506,3562,3588,3597,3627,3678,3703,3723,3738,3802,3841,3881,3900,4024,4067,4078,4085,4094,4112,4150,4203],"proposed,":[384],"which":[385,430,693,732,982,1001,1037,1171,1213,1238,1286,1327,1548,1787,1900,1956,2158,2245,2268,2806,3069,3124,3817],"employs":[386],"encoder-separator-decoder":[388,993,1064,1236],"architecture":[389,994,1395,1818,2282,2536],"implement":[391,474],"separation.":[393,850,915],"encoder":[395,1097],"can":[396,769,1002,1038,1083,1221,1287,1613,1719,2328,2393,2513,2638,2718,2855,3042,3062,3078,3090,3111,3158,3290,3377,3543,3769],"convert":[397],"signal":[399,413,547,811,831,849,914,1005,1101,1182,1296,1383,1462,1522,1541,1560,1592,1596,1603,1686,1703,1712,1832,1855,1907,1945,1981,2082,2218,2334,2641,2710,3374,3759,3799,3807,3865,3871],"into":[400,942,1102,1552,1859,1904,1982,1990,2249,2721,3174,3569,3599,3629,3884],"high-dimensional":[402,416,1068,1088,1104,1111,1124,1867,2017,2227,3716],"feature":[403,483,827,961,1112,1125,1210,1269,1325,1842,1894,2019,2164,2406,2465,3083,3181,3185,3213,3223,3271,3314,3339,3358,3442,3447,3467,3472,3512,3537,3580,3699,3728],"representation":[404,1843,1851,1879,1895,2020,3700,3711,3729],"separator":[407,1036,1108,1132,1862,2261,2285,2292,2343,2535],"achieve":[409,756,862,2832],"space.":[417],"core":[419,1135],"proposed":[422,499,686,904,991,1014,1061,1092,1233,1262,1335],"innovative":[428],"separator,":[429,1146,1827],"mainly":[432,601,2293],"composed":[433,3819],"attention-based":[435,442,1168,1272,2296,2306,2318,2323,2361,2428,2543,2562,2631,2672],"convolution":[436,443,449,1152,1169,1273,1930,1938,2011,2045,2053,2062,2070,2087,2113,2120,2297,2307,2319,2324,2362,2429,2447,2544,2563,2590,2632,2673,2725,2728,2736,2752,2766,2820,3235,3347,3438,3489,3744,3748],"units":[437],"residual":[439,2352,2377,2391],"connection.":[440],"unit":[444,909,2156,2298,2308,2363,2373,2381,2497,2512,2545,2564,2633,2674],"integrates":[445,2676],"large":[447,1150,2801],"kernel":[448,1151,3749],"modified":[451,1158,1283,1313,2154,2576,2625,2649,2682,3137,3152,3169],"global":[452,462,704,1159,1178,1292,2658,2690,2896,3088,3156,3218,3260,3566],"context":[453,1160,2804,3177,3195,3478],"(GC)":[454,1161],"block":[455,472,1154,1162,1196,1274,1281,1315,2309,2364,2430,2546,2565,2570,2578,2623,2627,2637,2651,2675,2679,2684,2717,2854,3041,3061,3077,3110,3142,3171],"simultaneously":[457,1174,1297],"capture":[458,1040,1175,1288,2212,2686,2756,3325],"local":[460,1176,1290,2688,2706,2758,3065],"information":[463,729,1179,1293,1572,2079,2516,2660,2691,2707,2760,3089,3157],"signal.":[466,1046,1130,2485,2508,2608,2670,2763,2814,3735,4125],"Furthermore,":[467],"we":[468,1147,1192,1255,1971,2619,3135,4009],"improve":[469,674,1184,1193,1299,2519],"GC":[471,1195,1284,1310,1314,2577,2626,2650,2683,3076,3109,3141,3170],"weighting":[476,479,1318,1322,3520,3541],"location":[478],"given":[482],"map,":[484,1326],"thus":[485,2224,2599,2662,3330],"further":[486,889,986,1039,3143],"enhancing":[487],"adaptability":[489,1330,2521,3127,3146],"model.":[492,1333,3149],"experimental":[494],"results":[495,3070],"show":[496],"that":[497,610,1191,1700,1989,2198,2355,2656,3086],"method":[500,917],"not":[501,587,1339,1966,3120,4095],"only":[502,770,870,1340,3063,3112],"outperforms":[503,1341],"mixed":[509,521,775,866,873,1347,1359,1510,1521,1540,1831,4193],"same":[513,779,1351,3618,3648,3747],"modulation,":[514,1480],"but":[515,1353,3119],"also":[516,836,1354,3882],"enables":[517,1172,1355,2072,2210,3323],"different":[524,1074,1076,1362,2231,3099,3460,4174],"modulations.":[525],"1":[526,1625,2442,2444,2449,2455,2457,2731,2733,2743,2745,2815,2817,2930,2937,2942,2947,2960,2967,2972,2977,3011,3013,3230,3232,3242,3244,3278,3280,3365,3367,3385,3404,3409,3411,3424,3426,3433,3435,3484,3486,3496,3498,3525],"INTRODUCTION":[527],"Single-channel":[528],"(SCBS)":[531],"technology":[532],"significance":[535],"fields":[538],"anti-interference,":[541],"paired":[542],"carrier":[543,1507],"multiple":[544,4163],"access":[545],"(PCMA)":[546],"monitoring":[548],"[1]":[549],"etc.":[550],"However,":[551,3058],"due":[552,792],"lack":[555],"prior":[557,2201],"knowledge,":[558],"it":[559,997,2696,3505],"has":[560,805,835],"been":[561,837],"challenge":[563],"solve":[565,1243,4004],"SCBS":[567,608,863,894,1245,2192],"problem":[568,899,1246,2197,2699],"[2].":[569],"Especially":[570],"signals,":[576,879,1078],"such":[580,813,4166],"as":[581,814,895,939,1252,1254,1364,1373,1616,1722,2300,2367,2386,2584,2829,2831,2858,3293,3375,3380,3546,3574,3576,3620,3650,3679,3754,3772,3843,3902,4044,4080,4114,4167,4187,4205],"successive":[582],"interference":[583],"cancellation":[584],"[3]":[585],"are":[586,592,600,787,1073,1509,2089,2097,2243,2246,2383,3020,3263,3664,3818,4042,4133,4184],"applicable":[588],"since":[589],"completely":[593],"overlap":[594],"time-frequency":[597,934],"domain.":[598,1010],"There":[599],"two":[602,901,1445,3345,3885],"types":[603,4165],"solutions":[605],"problem,":[609,4008],"is,":[611,2356],"model-driven":[614],"approaches":[615],"data-driven":[618],"approaches.":[619],"independent":[622],"component":[623,627],"analysis":[624,628],"sparse":[626],"poor":[631],"conditions":[636],"[4-7].":[637],"Particle":[638],"filtering":[639,1484],"(PF)":[640],"[8-10]":[641],"per-survivor":[643],"processing":[644,812,832],"(PSP)":[645],"[11-13]":[646],"algorithms":[647,786,968,976],"become":[649],"mainstream":[651],"tackle":[654],"SCBS.":[655],"Fynn":[656],"[9]":[657],"applied":[658,772,846,912,996,1241,2063,2754,3457],"idea":[660,715,2350],"Gibbs":[662],"sampling":[663],"PF":[665,678,691],"resampled":[667],"each":[668,1204,1404,1870,2360,3162,3165,3335,3463,3572],"particle":[669],"virtual":[671],"particles":[672,698],"parallelisation":[676],"while":[679,977,2084,2375,2643,2704],"reducing":[680],"accuracy.":[683],"Huo":[684],"[10]":[685],"gravitational":[688],"field":[689,841,2597,2703,2770],"resampling":[690],"algorithm,":[692],"increased":[694],"variability":[696],"enhanced":[700,1032],"search":[705],"removing":[707],"low-quality":[709],"particles.":[710],"Liao":[711],"[12]":[712],"adopted":[713,3842,4204],"soft-output":[718],"Viterbi":[719],"algorithm":[720,725,754,768,921,4030],"enable":[722],"PSP":[724,753,767,920],"output":[727,1557,2358,2425,2499,2838,3270,3376,3673],"soft":[728],"symbols,":[731],"achieved":[733,806,952],"better":[734],"performance.":[736,799,1186,1302],"To":[737,2609,4003,4153],"exploit":[738],"coding":[741],"information,":[742,1970],"Tu":[743],"[13]":[744],"combined":[745],"decoding":[747,760],"convolutional":[749,1016],"codes":[750],"joint":[757,897],"short":[762],"constraint":[763],"codes.":[764],"Unfortunately,":[765,3107],"be":[771,888,985,1084,1222,1614,1720,1958,2204,2719,2856,3091,3159,3291,3378,3544,3770,4154],"separate":[774],"modulation.":[780],"Moreover,":[781],"restricted":[788,3072],"practical":[790],"applications":[791],"recent":[801],"years,":[802],"learning":[804,844,2225,4069,4083,4148],"great":[807],"success":[808],"temporal":[810,830],"speech":[815,999,2065],"[14-16]":[817],"modulation":[819,1352,1363,1508,4164,4172],"recognition":[820],"[17],":[824],"demonstrating":[825],"powerful":[826],"extraction":[828],"capabilities.":[833],"It":[834,1187],"explored":[838],"Jiai":[851],"He":[852],"[18]":[853],"employed":[854,2822,3724],"three-layer":[856],"recurrent":[858,908],"neural":[859,947,1962,2591],"non-linear":[865,2495,2510],"signals.":[867,2233,2255,2280],"separated":[871,1085,1559,2241,2254,2279,3853,3870],"sine":[876],"cosine":[878,4201],"whose":[880,2095],"needs":[886,983],"tested.":[890],"[19]":[892],"modeled":[893],"detection":[898],"bidirectional":[906],"gated":[907],"(BGRU)":[910],"outperformed":[918],"terms":[923],"at":[927,3571],"low":[928,1532],"complexity.":[929],"Wei":[930],"[20]":[931],"fed":[932,1858,2492],"spectrum":[935],"sample":[937],"waveform":[938],"input":[940,2035,2369,3082,3511],"data":[941,962],"back":[944],"propagation":[945,1379,1442,1587,1597],"(BP)":[946],"network.":[948,1140,1992],"BP":[950],"bit":[956],"sequences":[957],"identifying":[959],"these":[964],"spectrums.":[965],"above":[967,2612],"reduced":[970],"computational":[971,3049],"compared":[973],"improved.":[987],"Luo":[988,1012],"[21]":[989],"first":[990,1240,1836,3822],"separation,":[1000],"model":[1003,1380,1443,1588,1598,2639,2824,3052,3064,3079,4037],"directly":[1006,1960,4185],"time":[1009,1801,2803],"Subsequently,":[1011],"successively":[1013],"time-domain":[1017,1100,1129,1906,2222,3734],"audio":[1018],"(Conv-TasNet)":[1021],"[22]":[1022],"Dual-path":[1024],"RNN":[1025],"(DPRNN)":[1026],"[23].":[1027],"improving":[1034],"long-range":[1042,2214,2666],"dependence":[1043,2215,2605,2667,2811],"Motivated":[1047],"Conv-TasNet,":[1049],"architecture.":[1065],"space,":[1069],"manifold":[1071,2236,2274],"distributions":[1072,2237,2275],"source":[1077,1081,1117,1461,1504,1662,1685,1702,1711,1854,1871,2232,2484,2507,3849,3857,4006,4189],"so":[1079,3085],"mask.":[1089],"network,":[1095,1400],"converts":[1098],"embedding":[1105,1841,2018],"representation,":[1106],"extracts":[1109],"mask":[1113,1126,1868,1886,2501,3717],"corresponding":[1114,1866,2229,2238,2276,2481,2504],"signal,":[1118],"decoder":[1121,3722,3737],"reverts":[1122],"design":[1143,2629],"integrate":[1148,2620],"(LKC)":[1153],"followed":[1155],"[24,":[1163],"25]":[1164],"construct":[1166,1864],"unit,":[1170],"worth":[1189],"noting":[1190],"assign":[1198,3459],"obtained":[1200],"weights":[1202,3097,3115,3461],"position":[1205,1321,3166,3464,3519,3540,3577],"map":[1211,2407,2466,3084,3214,3224,3272,3359,3473,3513,3581],"simultaneously,":[1212,2083],"facilitates":[1214],"task.":[1217],"Our":[1218],"key":[1219],"contributions":[1220],"summarised":[1223],"following:":[1226],"learning-based":[1230,2208],"approach":[1231,2209],"architecture,":[1237],"far":[1253],"know.":[1256],"Distinguishing":[1257],"from":[1258,2033,2423,3208,3476,3705],"algorithms,":[1260],"no":[1265],"longer":[1266],"requires":[1267,2199],"manual":[1268],"selection.":[1270],"An":[1271],"designed":[1276,1861,2345,3563],"LKC":[1280,2569,2622,2636,2678,2716,2853,3040,3060],"block,":[1285,2374],"We":[1303],"improved":[1305,4096],"structure":[1307,1921,3139],"block.":[1311,2320,2634],"implements":[1316],"along":[1319],"improves":[1328],"well.":[1365],"rest":[1367],"paper":[1370,4074],"structured":[1372],"follows.":[1374,3680,4045],"introduction":[1376],"provided":[1385],"Section":[1387,1389,1413,1428],"2;":[1388],"3":[1390,1803,3237,3305],"describes":[1391],"detail":[1393],"role":[1402],"component,":[1405],"training":[1408,3831,3845,3879,3889,3894,3898,4015],"model;":[1412],"4":[1414,3175,3491,3550,4101],"discusses":[1415],"effectiveness":[1417],"signals;":[1426],"Finally,":[1427,3796],"5":[1429],"summary":[1432],"paper.":[1435],"THE":[1437],"SIGNAL":[1438],"PROPAGATION":[1439],"MODEL":[1440],"depicted":[1451],"Figure":[1453,1814,2057,2289,2303,2587],"1.":[1454],"On":[1455,1516],"transmitter":[1457],"side,":[1458,1519],"baseband":[1460,1503,1539,3864,4182],"s":[1463,1638,1713,1723,3299,3906,3914,3916,3922,3933,3935,3941,3952,3954,3962,3971,3977,3979,3981,3989],"i":[1464,1469,1491,1562,1623,1628,1633,1637,1639,1655,1666,1672,1714,1724,1737,1739,1746,1755,1773,1796,1873,1888,1897,1909,3915,3934,3953],"(":[1465,1468,1493,1543,1563,1605,1618,1640,1645,1689,1715,1725,1774,1996,2000,2038,2114,2130,2139,2146,2167,2547,2779,2881,2890,2905,2912,2916,2926,2956,2987,3306,3386,3395,3551,3776,3787,3905,3919,3921,3926,3938,3940,3945],"t":[1466,1544,1564,1606,1619,1634,1641,1646,1690,1716,1726,1740,3302],")":[1467,1472,1545,1565,1607,1620,1642,1647,1691,1717,1727,1776,1998,2002,2132,2141,2148,2169,2787,2883,2892,2907,2914,2918,2928,2958,2989,3308,3397,3398,3553,3778,3791,3909,3923,3928,3929,3942,3947,3948],"=":[1470,1621,1624,1728,1731,1994,2128,2133,2170,2781,2860,2908,2929,2936,2941,2946,2959,2966,2971,2976,3295,3382,3548,3683,3774,3779,3910,3957,3973,3986],"1,2":[1471],"$\\begin{array}{ll}{s}_{i}(t)\\hfill":[1473],"&":[1474,2180,2186],"(i=1,2\\hfill":[1475],"\\end{array})$":[1476],"derived":[1478],"after":[1479,1506,1523,1575,1609,1948,3285],"upsampling,":[1481],"shaping":[1483,1790,4207],"symbol":[1487,1760,1763],"sequence":[1488],"k":[1490,1494,1730,1736,1742,1754],"${a}_{k}^{i}$":[1492,1756],"$k$":[1495],"indicates":[1496,1675,1928,1935,2041,2118,2895,3036],"index":[1498],"symbol).":[1501],"Two":[1502],"noise.":[1515,1698,4198],"receiver":[1518],"digital":[1524],"down":[1525],"conversion":[1526],"(DDC)":[1527],"passed":[1529,2433],"through":[1530,2434],"pass":[1533],"filter":[1534,1791,4202,4208],"(LPF)":[1535],"obtain":[1537,1891,2463,3253],"x":[1542,1604,1617,1946,1997,2001,2168,2171,2173,2181],"$x(t)$":[1546,1608],",":[1547,1899,2172,2783,2785,2951,2953,2981,2983,3005,3416,3623,3810,3907,3924,3943,3976],"subsequently":[1550],"sent":[1551],"y":[1561,3808,3908,3927,3946,3975,3987],"${y}_{i}(t)$":[1566],".":[1567,1875,2750,2789,3503,3584,3653,3720],"Lastly,":[1568],"code":[1570],"element":[1571,3669],"acquired":[1574],"matched":[1576],"filtering,":[1577],"sampling,":[1578],"judgment.":[1580],"FIGURE":[1581,1914,2103,2527,2537],"1Open":[1582],"figure":[1584,1917,2106,2530,2540],"viewerPowerPoint":[1585,1918,2107,2531,2541],"According":[1593],"described":[1599],"above,":[1600],"received":[1602,1944,1980,2422,3471],"DDC":[1610,1949],"LPF":[1612,1951],"expressed":[1615,1721,3379,3545,3771],"\u2211":[1622,1729,2934,2939,2944,2964,2969,2974],"h":[1627,1654,2940,2952,2970,2982],"e":[1629,3967,3984],"j":[1630],"\u0394":[1631,1664],"\u03c9":[1632,1665],"+":[1635,1643,2919,2923,3685,3688],"\u03b8":[1636,1671],"v":[1644,1688,3384,3393,3403,3418],"$x(t)=\\sum\\limits":[1648],"_{i=1}^{2}{h}_{i}{e}^{j\\left({\\Delta":[1649],"}{\\omega":[1650,1668],"}_{i}t+{\\theta":[1651],"}_{i}\\right)}{s}_{i}(t)+v(t)$":[1652],"(1)where":[1653],"${h}_{i}$":[1656],"denotes":[1657,1693,1757,1779,1799,2551,2557,3431],"amplitude":[1659,2076],"signal;":[1663,1771],"${\\Delta":[1667],"}_{i}$":[1669,1674,1798],"${\\theta":[1673],"frequency":[1677],"offset":[1678],"initial":[1680,4076],"phase":[1681,1969,2078],"ith":[1684,1710,1770],"respectively;":[1687],"$v(t)$":[1692],"additive":[1695],"white":[1696],"Gaussian":[1697],"Assuming":[1699],"MPSK":[1705,4169],"or":[1706],"MQAM":[1707,4171],"modulated,":[1708],"${s}_{i}(t)$":[1718],"\u2212":[1732,1741,1744,2910,2985,3911,3930,3988,4049],"\u221e":[1733,1734],"g":[1738,1772,2862,2878,2887,2902,3389],"T":[1743,1766],"\u03c4":[1745,1795],"${s}_{i}(t)=\\sum\\limits":[1747],"_{k=-\\infty":[1748],"}^{\\infty":[1749],"}{a}_{k}^{i}{g}_{i}\\left(t-kT-{\\tau":[1750],"}_{i}\\right)$":[1751],"(2)where":[1752],"kth":[1759],"period":[1764],"$T$":[1767],"\u22c5":[1775,2147,2891],"${g}_{i}(\\cdot":[1777],")$":[1778,2151,2894],"pulse":[1781],"response":[1782],"equivalent":[1784,4217],"filters,":[1786],"contains":[1788],"both":[1789,2687],"filter;":[1794],"${\\tau":[1797],"delay.":[1802],"DEEP":[1804],"SEPARATION":[1805],"NETWORK":[1806],"BASED":[1807],"ON":[1808],"ATTENTION":[1809],"MECHANISM":[1810],"As":[1811],"shown":[1812,2055,2287,2301,2585],"2,":[1815],"backbone":[1817,1920],"includes":[1823,3872],"three":[1824,2722],"sub-networks:":[1825],"encoder,":[1826],"decoder.":[1829,1913],"X":[1833,1993,2036,2131,2140],"$\\mathbf{X}$":[1834,2037],"converted":[1837,1903,2248,3803],"learnable":[1840,2098],"E":[1844,1880,2021,2127,2408,2467,2488,2882,2906,2909,2913,2917,2927,2949,2957,2979,2988,3674,3681,3712,3788],"$\\mathbf{E}$":[1845,1881,3713],"encoder.":[1848,3756],"encoded":[1850],"then":[1857,2247,2432,3258,4192],"M":[1872,1887,2502,3718,3790],"${\\mathbf{M}}_{i}$":[1874,1889],"Afterwards,":[1876],"coded":[1878,3710],"dot-producted":[1883],"masked":[1893,3698],"V":[1896,3701,3730,3777],"${\\mathbf{V}}_{i}$":[1898],"finally":[1902],"Y":[1908,3760,3773,3800,3827],"${\\mathbf{Y}}_{i}$":[1910],"2Open":[1915],"(2-D":[1926],"conv":[1927,1934,2136,2872,2876,3782],"2-D":[1929,1932,1936,2010,2052,2069,2112,3742],"operation;":[1931],"Trans":[1933],"transposed":[1937,3743],"operation)":[1939,2121],"Encoder":[1940],"Note":[1941],"that,":[1942,2487,3134],"$\\mathbf{x}$":[1947],"complex":[1954,3806,3867],"form,":[1955,3868],"cannot":[1957],"processed":[1959],"networks.":[1963,3057],"order":[1965],"lose":[1968],"splice":[1972],"real":[1974,3812,3873,3887,3920,3925],"imaginary":[1976,3814,3875,3892],"parts":[1977,3815],"two-dimensional":[1984],"(2-D)":[1985],"matrix":[1986,2552,3266,3521,3542],"before":[1987],"inputting":[1988],"Re":[1995],"Im":[1999],"$\\mathbf{X}=\\left[\\begin{array}{l}\\text{Re}(\\mathbf{x})\\hfill":[2003],"\\\\":[2004,2994],"\\mathrm{Im}(\\mathbf{x})\\hfill":[2005],"\\end{array}\\right]$":[2006],"(3)":[2007],"Then,":[2008],"utilised":[2013,2270,3093],"extract":[2015,2074,2100,2329,2340],"\u2208":[2022,2410,2450,2469,2739,2840,3007,3238,3274,3361,3405,3420,3492,3523,3604,3634,3761],"\u211d":[2023,2411,2451,2470,2740,2841,3008,3239,3275,3362,3406,3421,3493,3524,3605,3635,3762],"C":[2024,2039,2085,2092,2412,2453,2472,2741,2842,2932,2938,2962,2968,3009,3240,3276,3363,3407,3422,3494,3606,3636],"\u00d7":[2025,2027,2413,2415,2443,2454,2456,2473,2475,2732,2742,2744,2816,2843,2845,3010,3012,3231,3241,3243,3277,3279,3364,3366,3408,3410,3423,3425,3434,3485,3495,3497,3526,3528,3607,3609,3637,3639,3764],"H":[2026,2414,2474,2844,2933,2943,2963,2973,3527,3608,3638],"$\\mathbf{E}\\in":[2029],"{\\mathbb{R}}^{C\\times":[2030,2419,2747,2848,3017,3246,3282,3369,3413,3428,3500,3613,3643],"H\\times":[2031,2420,2479,2849,3532,3614,3644],"2}$":[2032,2421,2480,2850,3533,3615,3645],"$C$":[2040,2086,2093],"number":[2043,3046],"kernels).":[2046],"schematic":[2048,2109],"diagram":[2049,2110,2533],"3.":[2058],"Un-like":[2059],"1-D":[2061],"[22],":[2067],"operation":[2071,2124,3587],"kernels":[2088],"similar":[2090],"filters":[2094],"parameters":[2096,4038],"features.":[2102,2341],"3Open":[2104],"\u2217":[2115],"$\\ast":[2116],"$":[2117,2550,2556,3035,4041,4066],"specific":[2123],"process":[2125,2402],"encode":[2129],"\u210f":[2134,2145,2166,3387,3780],"W":[2135,2867,2873,3781],"d":[2138,3784],"$\\mathbf{E}=\\mathrm{encode}(\\mathbf{X})=\\mathrm{\\hslash":[2142],"}\\left({\\mathbf{W}}_{\\mathrm{conv}2\\mathrm{d}}(\\mathbf{X})\\right)$":[2143],"(4)where":[2144],"$\\hslash":[2149],"(\\cdot":[2150],"linear":[2155],"(ReLU),":[2157],"ensures":[2159],"non-negativity":[2161],"representation:":[2165],">":[2174,2182],"0":[2175],"0,":[2176],"otherwise":[2177],"$\\mathrm{\\hslash":[2178],"}(x)=\\left\\{\\begin{array}{c}\\begin{array}{ll}x,\\hfill":[2179],"0\\hfill":[2183],"\\end{array}\\\\":[2184],"\\begin{array}{ll}\\text{0,}\\hfill":[2185],"\\text{otherwise}\\hfill":[2187],"\\end{array}\\end{array}\\right.$":[2188],"(5)":[2189],"Separator":[2190],"mathematically":[2194],"ill-conditioned":[2196],"some":[2200],"knowledge":[2202],"addressed.":[2205],"modeling":[2220,3178,3196,3288,3479],"waveform,":[2223],"masks":[2228],"Thus":[2234],"obtained,":[2244],"one-dimensional":[2250],"waveforms":[2251],"paper,":[2258],"designed,":[2267],"generate":[2272],"4.":[2290],"consists":[2294,2566],"blocks":[2299,2325,2382],"5.":[2304,2588],"repeated":[2311],"R":[2312,2786],"$R$":[2313],"times":[2314],"form":[2316,3216],"N":[2321,2864,2880,2889,2904,3391],"$N$":[2322],"stacked":[2326],"together":[2327],"deep-level":[2330],"features":[2331,3207,3262,3568,3570],"enhance":[2336,3144],"ability":[2338],"motivated":[2347],"[26],":[2354],"used":[2366,4118,4129],"next":[2372],"paths":[2378],"all":[2380,3209],"summed":[2384],"up":[2385],"final":[2388],"output.":[2389],"This":[2390],"connection":[2392],"effectively":[2394,2663,3043],"prevent":[2395],"degradation":[2397],"accelerate":[2399],"convergence":[2401],"during":[2403],"training.":[2404],"\u2032":[2409,3603,3633,3656,3661,3675,3682,3687,3690],"${\\mathbf{E}}^{\\prime":[2417,3676,3691],"}\\in":[2418,3612,3642],"parametric":[2436],"activation":[2437,2496,2511,3354],"function":[2438,2899,3192,3251],"(PReLU)":[2439],"$1\\times":[2445,2734,2818,3233,3436,3487],"1$":[2446,2735,2819,3234,3437,3488],"U":[2448,2737,2865,3236,3304,3383,3392,3402,3417,3490,3549],"${\\mathbf{U}}_{\\mathbf{1}}\\in":[2458],"{\\mathbb{R}}^{2C\\times":[2459,2478],"1\\times":[2460,2748,3018,3247,3283,3370,3414,3429,3501],"1}$":[2461,2749,3019,3248,3284,3371,3415,3430,3502],"\u2033":[2468,2489],"${\\mathbf{E}}^{{\\prime\\prime}}\\in":[2477],"After":[2486,2835],"${\\mathbf{E}}^{{\\prime\\prime}}$":[2490],"$\\mathbf{M}$":[2503,3719],"control":[2514],"flow":[2517],"[27,":[2525],"28].":[2526],"4Open":[2528],"5Open":[2538],"\u2297":[2548,3297,3554],"$\\otimes":[2549],"multiplication,":[2553],"\u2295":[2554],"$\\oplus":[2555],"broadcast":[2558,3591],"element-wise":[2559,3592],"addition)":[2560],"(marked":[2571,2579,3198,3317,3450],"pink)":[2573],"yellow)":[2581],"series,":[2583],"Traditional":[2589],"networks":[2592],"limited":[2595,2701],"perceptual":[2596,2646,2702,2769],"difficulty":[2601],"capturing":[2602],"long-term":[2604,2810],"address":[2610],"issue,":[2613],"inspired":[2614],"literature":[2617],"[29],":[2618],"locally":[2642],"increasing":[2644],"field.":[2647],"models":[2657],"contextual":[2659,2759,3219,3261,3287,3567],"captures":[2664],"simultaneously.":[2692,3167],"way,":[2695],"overcomes":[2697],"neither":[2712],"lost.":[2713],"standard":[2715],"split":[2720],"parts:":[2723,3886],"depth":[2724],"(DW-Conv),":[2726],"depth-dilated":[2727],"(DW-D-Conv),":[2729],"${\\mathbf{U}}_{2}\\in":[2746],"Deep":[2751,2764],"dilation":[2765,2774],"expands":[2767],"setting":[2772],"factor":[2775,2792,4212],"r":[2777,2780,3918,3937,3956],"${2}^{r}$":[2778],"0,1,2":[2782],"\u22ef":[2784],"$(r=0,1,2,{{\\cdots}},R)$":[2788],"expansion":[2791],"set":[2794,4093],"grow":[2796],"exponentially":[2797],"ensure":[2799],"enough":[2802],"window,":[2805],"efficiently":[2807],"utilises":[2808],"correlation":[2826,3327],"between":[2827,3328,3847],"channels":[2828,3100,3118],"well":[2830,3575],"adaptation.":[2834],"normalisation,":[2836],"I":[2839,2859,3225,3296,3307,3514,3555,3582,3621,3651,3667,3684],"$\\mathbf{I}\\in":[2847],"written":[2857,3901],"f":[2861,2877,2886,2901,3298,3301,3388],"L":[2863,2879,2888,2903,3390,3765],"dw":[2868,2874],"-":[2869,2871,2875,3785],"D":[2870],"$\\mathbf{I}={f}_{gLN}\\left({\\mathbf{U}}_{2}\\left({\\mathbf{W}}_{\\mathrm{dw}\\mbox{-}\\mathrm{D}\\mbox{-}\\mathrm{conv}}\\left({\\mathbf{W}}_{\\mathrm{dw}\\mbox{-}\\mathrm{conv}}\\left({f}_{gLN}(\\mathbf{E})\\right)\\right)\\right)\\right)$":[2884],"(6)where":[2885],"${f}_{gLN}(\\cdot":[2893],"layer":[2897],"normalisation":[2898,3350],"[22]:":[2900],"Mean":[2911,2925,2986],"Var":[2915,2955],"\u03f5":[2920,3024],"\u2299":[2921,3033,3789],"\u03b3":[2922,3004],"\u03b2":[2924,3006],"c":[2935,2950,2965,2980],"w":[2945,2954,2975,2984],"$\\left\\{\\begin{array}{l}{f}_{gLN}(\\mathbf{E})=\\frac{\\mathbf{E}-\\text{Mean}(\\mathbf{E})}{\\sqrt{\\mathrm{Var}(\\mathbf{E})+{\\epsilon}}}\\odot":[2991],"\\gamma":[2992],"+\\beta":[2993],"\\text{Mean}(\\mathbf{E})=\\frac{1}{2CH}\\sum\\limits":[2995],"_{c=1}^{C}\\sum\\limits":[2996,3000],"_{h=1}^{H}\\sum\\limits":[2997,3001],"_{w=1}^{2}{\\mathbf{E}}_{c,h,w}\\\\":[2998],"\\mathrm{Var}(\\mathbf{E})=\\frac{1}{2CH}\\sum\\limits":[2999],"_{w=1}^{2}{\\left({\\mathbf{E}}_{c,h,w}-\\text{Mean}(\\mathbf{E})\\right)}^{2}\\end{array}\\right.$":[3002],"(7)where":[3003],"$\\gamma":[3014],",\\beta":[3015],"\\,\\in":[3016],"trainable":[3021],"parameters,":[3022,3048],"${\\epsilon}$":[3025],"small":[3028],"constant":[3029],"numerical":[3031],"stability;":[3032],"$\\odot":[3034],"dot-product.":[3038],"reduce":[3044],"cost,":[3050],"size":[3053,3619,3649,3750,3753],"building":[3055],"lightweight":[3056],"regions":[3066],"context,":[3068],"perception":[3073],"fields.":[3074],"whole":[3081],"fully":[3092],"rescale":[3095],"adjust":[3102,3113],"dependency":[3105],"[25].":[3106],"positions,":[3123],"limits":[3125],"module":[3130,3197,3289,3316,3341,3449,3480,3561],"[30].":[3131],"Based":[3132],"model,":[3153],"captured":[3155],"fused":[3160],"divided":[3173,3883],"sub-modules:":[3176],"module,":[3179,3183,3187,3444],"transformation":[3182,3186,3315,3340,3443,3448],"spatial":[3184,3446],"fusion":[3189,3560,3586],"module.":[3190],"green":[3201],"dotted":[3202,3321,3454],"box)":[3203,3322,3455],"aggregate":[3206],"locations":[3210],"feature.":[3220],"Firstly,":[3221],"$\\mathbf{I}$":[3226,3515,3583,3622,3652,3668],"subjected":[3228,3482],"${\\mathbf{U}}_{3}\\in":[3245],"softmax":[3250],"weights,":[3256],"calculated":[3264,3704,4025],"multiplication":[3267],"operation.":[3268],"O":[3273,3294,3396,3474,3552],"$\\mathbf{O}\\in":[3281],"represented":[3292],"o":[3300],"max":[3303],"$\\mathbf{O}=\\mathbf{I}\\otimes":[3309],"{f}_{soft\\mathrm{max}}\\left({\\mathbf{U}}_{3}(\\mathbf{I})\\right)$":[3310],"(8)":[3311],"blue":[3320],"channels,":[3329],"realising":[3331],"recalibration":[3333],"channel.":[3336],"implemented":[3343,3739],"bottleneck":[3346],"layers,":[3348],"function,":[3351],"function.":[3355],"$\\mathbf{A}\\in":[3368],"$\\mathbf{A}={\\mathbf{U}}_{v1}(\\mathrm{\\hslash":[3399],"}\\left({f}_{gLN}\\left({\\mathbf{U}}_{v2}(\\mathbf{O})\\right)\\right))$":[3400],"(9)where":[3401],"${\\mathbf{U}}_{v1}\\in":[3412],"${\\mathbf{U}}_{v2}\\in":[3427],"respectively.":[3439,3829],"Unlike":[3440],"purple":[3453],"map.":[3468,3538],"First,":[3469],"$\\mathbf{O}$":[3475],"${\\mathbf{U}}_{4}\\in":[3499],"Then":[3504,3654],"multiplied":[3507],"original":[3510,3536],"get":[3517],"G":[3522,3547,3625,3632,3660,3689],"$\\mathbf{G}\\in":[3530],"{\\mathbb{R}}^{1\\times":[3531],"$\\mathbf{G}={\\mathbf{U}}_{4}(\\mathbf{O})\\otimes":[3556],"\\mathbf{I}$":[3557],"(10)":[3558],"merge":[3565],"performed":[3589],"addition,":[3593],"i.e.,":[3594],"$\\mathbf{A}$":[3596],"expanded":[3598,3628],"tensor":[3601,3631],"${\\mathbf{A}}^{\\prime":[3611,3657],"$\\mathbf{G}$":[3626],"${\\mathbf{G}}^{\\prime":[3641,3662],"}$":[3658,3663,3677,3694,4062],"added":[3665],"element.":[3671],"}=\\mathbf{I}+{\\mathbf{A}}^{\\prime":[3692],"}+{\\mathbf{G}}^{\\prime":[3693],"(11)":[3695],"Decoder":[3696],"$\\mathbf{V}$":[3702,3731],"dot-product":[3707],"revert":[3726],"stride":[3752],"decoded":[3758,3798],"$\\mathbf{Y}\\in":[3766],"{\\mathbb{R}}^{2\\times":[3767],"L}$":[3768],"decode":[3775],"transpose":[3786],"$\\mathbf{Y}=\\mathrm{decode}(\\mathbf{V})=\\hslash":[3792],"\\left({\\mathbf{W}}_{\\mathrm{conv}2\\mathrm{d}\\mbox{-}\\mathrm{transpose}}(\\mathbf{E}\\odot":[3793],"\\mathbf{M})\\right)$":[3794],"(12)":[3795],"$\\mathbf{Y}$":[3801,3828],"$\\mathbf{y}$":[3809],"second":[3824],"rows":[3825],"Model":[3830],"general,":[3833],"negative":[3835],"minimum":[3836,4019],"scale-invariant":[3837],"source-to-noise":[3838],"ratio":[3839,4176],"(SI-SNR)":[3840],"loss":[3846,3880,3890,3903,3913,3932,3951,4020,4052],"[31].":[3859],"view":[3861],"parts,":[3876],"hence":[3877],"part":[3888,3893],"loss.":[3895,4035],"overall":[3897],"target":[3899,3963,3972,3990],"total":[3904,4053],"0.5":[3912,3931],"n":[3917,3936,3955],"imag":[3939,3944],"${\\text{loss}}_{\\text{total}}(\\mathbf{s},\\mathbf{y})=-0.5{\\text{loss}}_{sisnr}(\\text{real}(\\mathbf{s}),\\text{real}(\\mathbf{y}))-0.5{\\text{loss}}_{sisnr}(\\text{imag}(\\mathbf{s}),\\text{imag}(\\mathbf{y}))$":[3949],"(13)where,":[3950],"10":[3958,3960],"log":[3959],"\u2016":[3961,3964,3966,3969,3980,3982],"noise":[3968,3985],"\u3008":[3974],"\u3009":[3978],"$\\left\\{\\begin{array}{l}{\\text{loss}}_{sisnr}=10{\\mathrm{log}}_{10}\\frac{{\\Vert":[3991],"{\\mathbf{s}}_{\\text{target}}\\Vert":[3992],"}^{2}}{{\\Vert":[3993],"{\\mathrm{e}}_{\\text{noise}}\\Vert":[3994],"}^{2}}\\\\":[3995,4000],"{\\mathbf{s}}_{\\text{target}}=\\frac{\\langle":[3996],"\\mathbf{y},\\mathbf{s}\\rangle":[3997],"\\!\\!\\mathbf{s}}{{\\Vert":[3998],"\\mathbf{s}\\Vert":[3999],"{\\mathbf{e}}_{\\text{noise}}=\\mathbf{y}-{\\mathbf{s}}_{\\text{target}}\\end{array}\\right.$":[4001],"(14)":[4002],"permutation":[4007,4013],"use":[4010],"utterance-level":[4012],"invariant":[4014],"calculate":[4017],"[32].":[4021],"gradient":[4023],"Adam":[4028],"optimisation":[4029],"[33]":[4031],"\u03d5":[4039,4046,4048,4055],"$\\phi":[4040,4056],"updated":[4043],"\u2190":[4047],"\u03b7":[4050,4064],"\u2202":[4051,4054],"{\\leftarrow}\\phi":[4057],"-\\eta":[4058],"\\frac{\\partial":[4059],"{\\text{loss}}_{\\text{total}}}{\\partial":[4060],"\\phi":[4061],"(15)where":[4063],"$\\eta":[4065],"rate,":[4070],"value":[4077],"taken":[4079],"0.0003.":[4081],"rate":[4084],"halved":[4086],"if":[4087],"accuracy":[4089],"validation":[4092],"consecutive":[4099],"epochs.":[4100],"EXPERIMENTAL":[4102],"RESULTS":[4103],"AND":[4104],"ANALYSIS":[4105],"simulation":[4108,4116],"experiments,":[4109],"Matlab":[4110],"2019":[4111],"chosen":[4113],"platform":[4117],"generation":[4121],"hardware":[4127],"resources":[4128],"experiment":[4132],"RTX":[4134],"3090":[4135],"GPU,":[4136],"Intel":[4137],"Xeon":[4138],"Silver":[4139],"4210R":[4140],"2.40":[4142],"GHz":[4143],"CPU,":[4144],"framework":[4149],"PyTorch":[4151],"1.10.":[4152],"convincing,":[4156],"experiments":[4159],"involve":[4160],"BPSK,":[4168],"signal-to-noise":[4175],"(SNR)":[4177],"10\u201322":[4179],"dB.":[4180],"generated":[4186],"raised":[4200],"roll-off":[4211],"0.35":[4214],"mem":[4219]},"cited_by_api_url":"https://api.openalex.org/works?filter=cites:W4308443027","counts_by_year":[{"year":2023,"cited_by_count":2}],"updated_date":"2024-11-26T20:54:34.767364","created_date":"2022-11-11"}