Publication Information

{'id': 'https://openalex.org/W4393304331', 'doi': 'https://doi.org/10.31274/td-20240329-380', 'title': 'Hardware-aware design, search, and optimization of deep neural networks', 'display_name': 'Hardware-aware design, search, and optimization of deep neural networks', 'publication_year': 2023, 'publication_date': '2023-01-01', 'ids': {'openalex': 'https://openalex.org/W4393304331', 'doi': 'https://doi.org/10.31274/td-20240329-380'}, 'language': 'en', 'primary_location': {'is_oa': True, 'landing_page_url': 'https://doi.org/10.31274/td-20240329-380', 'pdf_url': 'https://dr.lib.iastate.edu/bitstreams/dba34f50-99cd-4693-8559-0ce00f74f3a1/download', 'source': None, 'license': None, 'license_id': None, 'version': 'publishedVersion', 'is_accepted': True, 'is_published': True}, 'type': 'dissertation', 'type_crossref': 'dissertation', 'indexed_in': ['crossref'], 'open_access': {'is_oa': True, 'oa_status': 'bronze', 'oa_url': 'https://dr.lib.iastate.edu/bitstreams/dba34f50-99cd-4693-8559-0ce00f74f3a1/download', 'any_repository_has_fulltext': True}, 'authorships': [{'author_position': 'first', 'author': {'id': 'https://openalex.org/A5094275716', 'display_name': 'Sai Subra Chitty-Venkata', 'orcid': None}, 'institutions': [{'id': 'https://openalex.org/I173911158', 'display_name': 'Iowa State University', 'ror': 'https://ror.org/04rswrd78', 'country_code': 'US', 'type': 'education', 'lineage': ['https://openalex.org/I173911158']}], 'countries': ['US'], 'is_corresponding': True, 'raw_author_name': 'Sai Subra Chitty-Venkata', 'raw_affiliation_strings': ['Iowa State University Ames, Iowa 2023'], 'affiliations': [{'raw_affiliation_string': 'Iowa State University Ames, Iowa 2023', 'institution_ids': ['https://openalex.org/I173911158']}]}], 'institution_assertions': [], 'countries_distinct_count': 1, 'institutions_distinct_count': 1, 'corresponding_author_ids': ['https://openalex.org/A5094275716'], 'corresponding_institution_ids': ['https://openalex.org/I173911158'], 'apc_list': None, 'apc_paid': None, 'fwci': None, 'has_fulltext': True, 'fulltext_origin': 'pdf', 'cited_by_count': 0, 'citation_normalized_percentile': {'value': 0.0, 'is_in_top_1_percent': False, 'is_in_top_10_percent': False}, 'cited_by_percentile_year': {'min': 0, 'max': 67}, 'biblio': {'volume': None, 'issue': None, 'first_page': None, 'last_page': None}, 'is_retracted': False, 'is_paratext': False, 'primary_topic': {'id': 'https://openalex.org/T10036', 'display_name': 'Advanced Neural Network Applications', 'score': 0.9997, 'subfield': {'id': 'https://openalex.org/subfields/1707', 'display_name': 'Computer Vision and Pattern Recognition'}, '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/T10036', 'display_name': 'Advanced Neural Network Applications', 'score': 0.9997, 'subfield': {'id': 'https://openalex.org/subfields/1707', 'display_name': 'Computer Vision and Pattern Recognition'}, '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/T12535', 'display_name': 'Machine Learning and Data Classification', 'score': 0.9922, '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'}}, {'id': 'https://openalex.org/T11689', 'display_name': 'Adversarial Robustness in Machine Learning', 'score': 0.991, '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': [], 'concepts': [{'id': 'https://openalex.org/C41008148', 'wikidata': 'https://www.wikidata.org/wiki/Q21198', 'display_name': 'Computer science', 'level': 0, 'score': 0.76862246}, {'id': 'https://openalex.org/C81363708', 'wikidata': 'https://www.wikidata.org/wiki/Q17084460', 'display_name': 'Convolutional neural network', 'level': 2, 'score': 0.57617706}, {'id': 'https://openalex.org/C50644808', 'wikidata': 'https://www.wikidata.org/wiki/Q192776', 'display_name': 'Artificial neural network', 'level': 2, 'score': 0.5686109}, {'id': 'https://openalex.org/C108583219', 'wikidata': 'https://www.wikidata.org/wiki/Q197536', 'display_name': 'Deep learning', 'level': 2, 'score': 0.5671857}, {'id': 'https://openalex.org/C154945302', 'wikidata': 'https://www.wikidata.org/wiki/Q11660', 'display_name': 'Artificial intelligence', 'level': 1, 'score': 0.55650055}, {'id': 'https://openalex.org/C118524514', 'wikidata': 'https://www.wikidata.org/wiki/Q173212', 'display_name': 'Computer architecture', 'level': 1, 'score': 0.45268267}, {'id': 'https://openalex.org/C147168706', 'wikidata': 'https://www.wikidata.org/wiki/Q1457734', 'display_name': 'Recurrent neural network', 'level': 3, 'score': 0.4519784}, {'id': 'https://openalex.org/C119857082', 'wikidata': 'https://www.wikidata.org/wiki/Q2539', 'display_name': 'Machine learning', 'level': 1, 'score': 0.43794465}, {'id': 'https://openalex.org/C175202392', 'wikidata': 'https://www.wikidata.org/wiki/Q2434543', 'display_name': 'Time delay neural network', 'level': 3, 'score': 0.4243533}, {'id': 'https://openalex.org/C45374587', 'wikidata': 'https://www.wikidata.org/wiki/Q12525525', 'display_name': 'Computation', 'level': 2, 'score': 0.411336}, {'id': 'https://openalex.org/C113775141', 'wikidata': 'https://www.wikidata.org/wiki/Q428691', 'display_name': 'Computer engineering', 'level': 1, 'score': 0.3690521}, {'id': 'https://openalex.org/C199360897', 'wikidata': 'https://www.wikidata.org/wiki/Q9143', 'display_name': 'Programming language', 'level': 1, 'score': 0.110364705}], 'mesh': [], 'locations_count': 1, 'locations': [{'is_oa': True, 'landing_page_url': 'https://doi.org/10.31274/td-20240329-380', 'pdf_url': 'https://dr.lib.iastate.edu/bitstreams/dba34f50-99cd-4693-8559-0ce00f74f3a1/download', 'source': None, 'license': None, 'license_id': None, 'version': 'publishedVersion', 'is_accepted': True, 'is_published': True}], 'best_oa_location': {'is_oa': True, 'landing_page_url': 'https://doi.org/10.31274/td-20240329-380', 'pdf_url': 'https://dr.lib.iastate.edu/bitstreams/dba34f50-99cd-4693-8559-0ce00f74f3a1/download', 'source': None, 'license': None, 'license_id': None, 'version': 'publishedVersion', 'is_accepted': True, 'is_published': True}, 'sustainable_development_goals': [], 'grants': [], 'datasets': [], 'versions': [], 'referenced_works_count': 117, 'referenced_works': ['https://openalex.org/W1686810756', 'https://openalex.org/W1902934009', 'https://openalex.org/W2017369466', 'https://openalex.org/W2096443114', 'https://openalex.org/W2097117768', 'https://openalex.org/W2108598243', 'https://openalex.org/W2112796928', 'https://openalex.org/W2163605009', 'https://openalex.org/W2194775991', 'https://openalex.org/W2276892413', 'https://openalex.org/W2278433511', 'https://openalex.org/W2300242332', 'https://openalex.org/W2469490737', 'https://openalex.org/W2531409750', 'https://openalex.org/W2547875792', 'https://openalex.org/W2553303224', 'https://openalex.org/W2612076670', 'https://openalex.org/W2707890836', 'https://openalex.org/W27224072', 'https://openalex.org/W2745868649', 'https://openalex.org/W2767260595', 'https://openalex.org/W2776940252', 'https://openalex.org/W2784372305', 'https://openalex.org/W2798724095', 'https://openalex.org/W2809624076', 'https://openalex.org/W2810271953', 'https://openalex.org/W2810704618', 'https://openalex.org/W2883780447', 'https://openalex.org/W2885820039', 'https://openalex.org/W2886438137', 'https://openalex.org/W2903260438', 'https://openalex.org/W2909929243', 'https://openalex.org/W2911491685', 'https://openalex.org/W2938555618', 'https://openalex.org/W2951104886', 'https://openalex.org/W2951245151', 'https://openalex.org/W2953212265', 'https://openalex.org/W2953384591', 'https://openalex.org/W2955425717', 'https://openalex.org/W2962746461', 'https://openalex.org/W2962965870', 'https://openalex.org/W2963037463', 'https://openalex.org/W2963122961', 'https://openalex.org/W2963163009', 'https://openalex.org/W2963367920', 'https://openalex.org/W2963372104', 'https://openalex.org/W2963821229', 'https://openalex.org/W2963918968', 'https://openalex.org/W2963989532', 'https://openalex.org/W2964108906', 'https://openalex.org/W2964259004', 'https://openalex.org/W2964515685', 'https://openalex.org/W2967733054', 'https://openalex.org/W2969414999', 'https://openalex.org/W2971882152', 'https://openalex.org/W2980137827', 'https://openalex.org/W2981406437', 'https://openalex.org/W2981701755', 'https://openalex.org/W2982479999', 'https://openalex.org/W2991907795', 'https://openalex.org/W2994749257', 'https://openalex.org/W2997958863', 'https://openalex.org/W3010249414', 'https://openalex.org/W3013854788', 'https://openalex.org/W3016142271', 'https://openalex.org/W3016542674', 'https://openalex.org/W3020881633', 'https://openalex.org/W3021575660', 'https://openalex.org/W3024621361', 'https://openalex.org/W3030728803', 'https://openalex.org/W3034368386', 'https://openalex.org/W3034887213', 'https://openalex.org/W3035130950', 'https://openalex.org/W3035251378', 'https://openalex.org/W3038988173', 'https://openalex.org/W3044604993', 'https://openalex.org/W3046770517', 'https://openalex.org/W3080443832', 'https://openalex.org/W3081305497', 'https://openalex.org/W3091602279', 'https://openalex.org/W3095959945', 'https://openalex.org/W3097528158', 'https://openalex.org/W3118608800', 'https://openalex.org/W3121907463', 'https://openalex.org/W3121924028', 'https://openalex.org/W3127067080', 'https://openalex.org/W3130554079', 'https://openalex.org/W3131920484', 'https://openalex.org/W3137264163', 'https://openalex.org/W3145444543', 'https://openalex.org/W3154028478', 'https://openalex.org/W3160878819', 'https://openalex.org/W3166395393', 'https://openalex.org/W3168376295', 'https://openalex.org/W3176537956', 'https://openalex.org/W3193559583', 'https://openalex.org/W3195426209', 'https://openalex.org/W3211833531', 'https://openalex.org/W4223527846', 'https://openalex.org/W4230563027', 'https://openalex.org/W4239385313', 'https://openalex.org/W4247198796', 'https://openalex.org/W4281734992', 'https://openalex.org/W4283368428', 'https://openalex.org/W4287208846', 'https://openalex.org/W4287257983', 'https://openalex.org/W4287639545', 'https://openalex.org/W4287715829', 'https://openalex.org/W4287907702', 'https://openalex.org/W4297775537', 'https://openalex.org/W4297778814', 'https://openalex.org/W4297813615', 'https://openalex.org/W4297917482', 'https://openalex.org/W4309845474', 'https://openalex.org/W4312257891', 'https://openalex.org/W4375869345', 'https://openalex.org/W4394650345'], 'related_works': ['https://openalex.org/W4298287631', 'https://openalex.org/W3206636855', 'https://openalex.org/W2953061907', 'https://openalex.org/W2890297197', 'https://openalex.org/W2373874059', 'https://openalex.org/W2109916967', 'https://openalex.org/W2101697354', 'https://openalex.org/W1847088711', 'https://openalex.org/W1538606284', 'https://openalex.org/W1538193578'], 'abstract_inverted_index': {'Deep': [0, 78], 'Learning': [1, 165], 'has': [2, 131, 390, 609], 'achieved': [3, 557], 'remarkable': [4], 'progress': [5], 'in': [6, 121, 127, 145, 242, 257, 278, 394, 414, 444, 452, 457, 621, 654, 695], 'the': [7, 75, 106, 113, 119, 152, 172, 206, 236, 268, 363, 374, 435, 445, 473, 485, 497, 506, 512, 544, 548, 554, 566, 613, 622, 660, 675, 679, 683, 687, 745, 752, 762, 781, 815, 841, 853, 881], 'last': [8], 'decade': [9], 'due': [10, 475, 598], 'to': [11, 41, 45, 73, 89, 112, 139, 170, 196, 204, 226, 306, 347, 381, 418, 467, 476, 495, 522, 592, 599, 631, 721, 742, 805, 834, 880], 'its': [12, 128, 618], 'powerful': [13], 'automatic': [14], 'representation': [15], 'capability': [16], 'for': [17, 199, 228, 280, 313, 533, 690, 764, 807, 837], 'a': [18, 141, 160, 168, 180, 219, 316, 369, 401, 407, 411, 421, 441, 639, 648, 665, 723, 748, 768, 787, 802], 'variety': [19, 317], 'of': [20, 68, 77, 109, 116, 162, 175, 221, 244, 318, 368, 400, 484, 511, 515, 547, 561, 573, 607, 677, 710, 736, 747, 786, 843, 848, 855], 'tasks,': [21, 130, 289], 'such': [22, 49, 86, 92, 101, 260, 290, 320, 469, 777], 'as': [23, 50, 87, 93, 102, 148, 261, 291, 321, 398, 416], 'Image': [24, 292], 'Recognition,': [25, 27], 'Speech': [26], 'and': [28, 63, 154, 183, 210, 264, 272, 300, 310, 325, 365, 396, 439, 454, 459, 538, 563, 576, 583, 617, 663, 667, 681, 716, 726, 739, 798, 810, 819, 826, 830, 840, 845, 873], 'Machine': [29, 164, 296], 'Translation.': [30], 'This': [31], 'success': [32], 'is': [33, 39, 71, 136, 167, 192, 218, 225, 423, 479, 633], 'associated': [34], 'with': [35, 336, 410, 629, 638, 674, 761], 'network': [36, 823], 'design,': [37], 'which': [38, 125, 354, 655], 'crucial': [40], 'feature': [42], 'representation,': [43], 'leading': [44, 521], 'many': [46, 288], 'innovative': [47], 'architectures': [48, 462], 'Convolutional': [51], 'Neural': [52, 56, 60, 79, 156, 176, 214, 343, 771], 'Network': [53, 57, 61, 177, 344, 371], '(CNN),': [54], 'Recurrent': [55], '(RNN),': [58], 'Graph': [59], '(GNN)': [62], 'Transformers.': [64], 'A': [65, 605], 'wide': [66], 'range': [67], 'hardware': [69, 85], 'platforms': [70], 'available': [72], 'accelerate': [74, 468], 'performance': [76, 263, 286, 341, 727], 'Networks': [80], '(DNNs),': [81], 'ranging': [82], 'from': [83], 'general-purpose': [84], 'CPUs': [88], 'special-purpose': [90], 'devices': [91, 319], 'Tensor': [94, 455, 518, 539, 580, 693, 731, 860], 'Processing': [95], 'Unit': [96], '(TPU).': [97], 'High-performance': [98], 'computing': [99], 'systems': [100, 588], 'GPUs': [103, 541, 864], 'effectively': [104], 'reduce': [105], 'computation': [107, 211], 'time': [108, 689], 'DNNs.': [110], 'Due': [111], 'slowing': [114], 'down': [115], "Moore's": [117], 'law,': [118], 'research': [120], 'developing': [122], 'Domain-Specific': [123], 'Hardware,': [124], 'excels': [126], 'assigned': [129], 'gained': [132], 'significance.': [133], 'Therefore,': [134, 472], 'it': [135, 149], 'not': [137, 232, 356, 464, 504, 634], 'straightforward': [138], 'choose': [140], 'platform': [142], 'that': [143, 230, 283], 'works': [144], 'all': [146], 'scenarios,': [147], 'depends': [150], 'on': [151, 179, 235, 267, 287, 315, 578, 612, 659, 862, 871], 'application': [153], 'environment.': [155], 'Architecture': [157, 215, 772], 'Search': [158, 216, 773, 800, 832], '(NAS),': [159], 'subset': [161], 'Automatic': [163], '(AutoML),': [166], 'method': [169, 191, 433, 532, 556, 568], 'automate': [171], 'design': [173, 307, 783], 'process': [174, 784], 'architecture': [178, 203, 782], 'given': [181, 237], 'task': [182, 340], 'dataset': [184, 238], 'without': [185], 'significant': [186], 'human': [187], 'intervention.': [188], 'The': [189, 249, 448, 550, 729], 'NAS': [190, 271, 803], 'an': [193, 200, 558, 570], 'intelligent': [194], 'algorithm': [195, 673], 'automatically': [197], 'search': [198, 227, 806, 817, 836], 'efficient': [201, 281, 691, 808, 867], 'neural': [202, 822], 'save': [205], "researcher's": [207], 'manual': [208], 'effort': [209], 'time.': [212], 'Hardware-aware': [213], '(HW-NAS)': [217], 'class': [220], 'problems': [222, 778], 'whose': [223], 'goal': [224], 'networks': [229, 256, 332, 552, 870], 'are': [231, 302, 463, 589, 714], 'only': [233, 574], 'accurate': [234], 'but': [239], 'also': [240, 733], 'hardware-efficient': [241], 'terms': [243], 'latency,': [245], 'energy,': [246], 'size,': [247], 'etc.': [248, 298], 'resulting': [250], 'searched': [251], 'models': [252, 282, 335, 813, 878], 'outperform': [253], 'manually': [254, 882], 'designed': [255, 466, 883], 'several': [258, 488], 'aspects,': [259], 'model': [262, 359, 376, 402, 422, 437, 650], 'inference': [265, 314, 399, 746], 'latency': [266], 'actual': [269], 'hardware.': [270], 'HW-NAS': [273], 'have': [274, 492], 'been': [275, 493, 610], 'very': [276], 'successful': [277], 'searching': [279, 866], 'achieve': [284, 722], 'State-of-the-art': [285], 'Classification,': [293], 'Object': [294], 'Detection,': [295], 'Translation,': [297], 'Pruning': [299, 345, 429, 530, 671], 'Quantization': [301, 372, 388], 'two': [303], 'important': [304], 'techniques': [305], 'lightweight,': [308], 'memory-efficient,': [309], 'hardware-friendly': [311], 'methods': [312, 328, 491, 502, 857], 'CPU,': [322], 'GPU,': [323], 'ASIC,': [324], 'FPGA.': [326], 'These': [327], 'successfully': [329], 'compressed': [330], 'large': [331], 'into': [333], 'smaller': [334], 'negligible': [337, 412, 480], 'accuracy': [338, 725], 'or': [339, 350, 430, 517, 595, 602, 757], 'loss.': [342], 'refers': [346], 'removing': [348, 482, 682], 'redundant': [349], 'unimportant': [351], 'weights/nodes/neurons/filters': [352], 'parameters': [353], 'do': [355, 503], 'significantly': [357, 438], 'hinder': [358], 'performance,': [360], 'thereby': [361], 'reducing': [362], 'size': [364, 669, 839], 'computational': [366], 'complexity': [367], 'model.': [370, 641, 749], 'converts': [373], 'high-precision': [375], 'weights/parameters': [377], '(Floating': [378], 'point': [379], '32)': [380], 'low': [382, 408], 'precision': [383, 409, 763, 869], '(Integer': [384], '8,': [385], 'Integer': [386], '4).': [387], 'methodology': [389], 'attracted': [391], 'much': [392], 'attention': [393], 'academia': [395], 'industry': [397], 'can': [403, 657, 775], 'be': [404], 'performed': [405], 'at': [406, 425, 686], 'drop': [413], 'accuracy,': [415], 'opposed': [417], 'training': [419], 'where': [420, 707], 'trained': [424], 'high': [426], 'precision.': [427], 'Weight': [428], 'element-wise': [431], 'pruning': [432, 478, 490, 706], 'shrinks': [434], 'DNN': [436], 'introduces': [440], 'considerable': [442], 'sparsity': [443, 498], 'weight': [446, 477], 'matrices.': [447, 471], 'uniform': [449], 'systolic': [450, 614, 661], 'arrays': [451], 'TPU': [453], 'Cores': [456, 694], 'Volta': [458], 'Turing': [460, 579], 'GPU': [461, 582, 700], 'explicitly': [465], 'sparse': [470, 705, 809], 'speedup': [474, 560, 572], 'despite': [481], '90\\%': [483], 'parameters.': [486], 'Later,': [487], 'node': [489], 'developed': [494], 'resolve': [496], 'bottlenecks.': [499], 'However,': [500, 624], 'these': [501], 'consider': [505], 'underlying': [507, 545], 'Hardware': [508, 527, 587], 'dimension': [509, 546], '(size': [510], 'array,': [513, 662], 'number': [514, 842], 'CPUs)': [516], 'Core': [519, 581, 732, 861], 'precision,': [520], 'suboptimal': [523], 'performance.': [524], 'We': [525, 851], 'develop': [526], 'Dimension': [528], 'Aware': [529], '(HDAP)': [531], 'array-based': [534], 'accelerators,': [535], 'multi-core': [536], 'CPUs,': [537], 'Core-enabled': [540], 'by': [542, 779, 865], 'considering': [543], 'system.': [549], 'node-pruned': [551], 'using': [553], 'HDAP': [555], 'average': [559, 571], '3.2x': [562], '4.2x,': [564], 'whereas': [565], 'baseline': [567], 'attained': [569], '1.5x': [575], '1.6x': [577], 'Eyeriss': [584], 'architecture,': [585], 'respectively.': [586], 'often': [590], 'prone': [591], 'soft': [593], 'errors': [594], 'permanent': [596], 'faults': [597, 656, 680], 'external': [600], 'conditions': [601], 'internal': [603, 684], 'scaling.': [604], 'lot': [606], 'work': [608], 'done': [611], 'array': [615, 668], 'implementation': [616], 'reliability': [619], 'concerns': [620], 'past.': [623], 'their': [625], 'fault': [626, 640, 649], 'tolerance': [627], 'perspective': [628], 'respect': [630], 'DNNs': [632], 'yet': [635], 'fully': [636], 'understood': [637], 'In': [642, 790], 'our': [643, 856], 'work,': [644, 792], 'we': [645, 793], 'first': [646], 'present': [647], 'i.e.,': [651], 'different': [652], 'sequences': [653], 'occur': [658], 'co-design': [664], 'fault-based': [666], 'based': [670], '(FPAP)': [672], 'intent': [676], 'bypassing': [678], 'redundancy': [685], 'same': [688], 'inference.': [692], 'Nvidia': [696, 863], 'Ampere': [697], '100': [698], '(A100)': [699], 'support': [701], '(1)': [702], '2:4': [703, 758], 'fine-grained': [704], '2': [708], 'out': [709], 'every': [711], '4': [712], 'elements': [713], 'pruned': [715], '(2)': [717], 'traditional': [718], 'dense': [719], 'multiplication': [720, 741], 'good': [724], 'trade-off.': [728], 'A100': [730, 859], 'takes': [734], 'advantage': [735], '1-bit,': [737], '4-bit,': [738], '8-bit': [740], 'speed': [743], 'up': [744], 'Hence,': [750], 'finding': [751], 'right': [753], 'matrix': [754], 'type': [755], '(dense': [756], 'sparse)': [759], 'along': [760], 'each': [765, 849], 'layer': [766], 'becomes': [767], 'combinatorial': [769], 'problem.': [770], '(NAS)': [774], 'alleviate': [776], 'automating': [780], 'instead': [785], 'brute-force': [788], 'search.': [789], 'this': [791], 'propose': [794], '\\textbf{(i)}': [795], 'Mixed': [796], 'Sparse': [797, 829, 885], 'Precision': [799, 831], '(MSPS),': [801], 'framework': [804], 'mixed-precision': [811], 'quantized': [812], 'within': [814], 'predefined': [816], 'space': [818], 'fixed': [820], 'backbone': [821], '(Eg.': [824], 'ResNet50),': [825], '\\textbf{(ii)}': [827], 'Architecture,': [828], '(ASPS)': [833], 'jointly': [835], 'kernel': [838], 'filters,': [844], 'sparse-precision': [846], 'combination': [847], 'layer.': [850], 'illustrate': [852], 'effectiveness': [854], 'targeting': [858], 'sparse-mixed': [868], 'ResNet50': [872], 'achieving': [874], 'better': [875], 'accuracy-latency': [876], 'trade-off': [877], 'compared': [879], 'Uniform': [884], 'Int8': [886], 'networks.': [887]}, 'cited_by_api_url': 'https://api.openalex.org/works?filter=cites:W4393304331', 'counts_by_year': [], 'updated_date': '2025-01-05T05:18:04.305807', 'created_date': '2024-03-30'}