%PDF-1.3 1 0 obj << /Kids [ 4 0 R 5 0 R 6 0 R 7 0 R 8 0 R 9 0 R 10 0 R 11 0 R 12 0 R 13 0 R 14 0 R ] /Type /Pages /Count 11 >> endobj 2 0 obj << /Subject (Neural Information Processing Systems http\072\057\057nips\056cc\057) /Publisher (Curran Associates\054 Inc\056) /Language (en\055US) /Created (2019) /EventType (Poster) /Description-Abstract (Autoregressive sequence models achieve state\055of\055the\055art performance in domains like machine translation\056 However\054 due to the autoregressive factorization nature\054 these models suffer from heavy latency during inference\056 Recently\054 non\055autoregressive sequence models were proposed to speed up the inference time\056 However\054 these models assume that the decoding process of each token is conditionally independent of others\056 Such a generation process sometimes makes the output sentence inconsistent\054 and thus the learned non\055autoregressive models could only achieve inferior accuracy compared to their autoregressive counterparts\056 To improve then decoding consistency and reduce the inference cost at the same time\054 we propose to incorporate a structured inference module into the non\055autoregressive models\056 Specifically\054 we design an efficient approximation for Conditional Random Fields \050CRF\051 for non\055autoregressive sequence models\054 and further propose a dynamic transition technique to model positional contexts in the CRF\056 Experiments in machine translation show that while increasing little latency \0508\17614ms\054 our model could achieve significantly better translation performance than previous non\055autoregressive models on different translation datasets\056 In particular\054 for the WMT14 En\055De dataset\054 our model obtains a BLEU score of 26\05680\054 which largely outperforms the previous non\055autoregressive baselines and is only 0\05661 lower in BLEU than purely autoregressive models\056) /Producer (PyPDF2) /Title (Fast Structured Decoding for Sequence Models) /Date (2019) /ModDate (D\07220200212223513\05508\04700\047) /Published (2019) /Type (Conference Proceedings) /firstpage (3016) /Book (Advances in Neural Information Processing Systems 32) /Description (Paper accepted and presented at the Neural Information Processing Systems Conference \050http\072\057\057nips\056cc\057\051) /Editors (H\056 Wallach and H\056 Larochelle and A\056 Beygelzimer and F\056 d\047Alch\351\055Buc and E\056 Fox and R\056 Garnett) /Author (Zhiqing Sun\054 Zhuohan Li\054 Haoqing Wang\054 Di He\054 Zi Lin\054 Zhihong Deng) /lastpage (3026) >> endobj 3 0 obj << /Type /Catalog /Pages 1 0 R >> endobj 4 0 obj << /Contents 15 0 R /Parent 1 0 R /Resources 16 0 R /MediaBox [ 0 0 612 792 ] /Annots [ 75 0 R 76 0 R 77 0 R 78 0 R 79 0 R 80 0 R 81 0 R 82 0 R 83 0 R 84 0 R 85 0 R ] /Type /Page >> endobj 5 0 obj << /Contents 86 0 R /Parent 1 0 R /Resources 87 0 R /Group 97 0 R /MediaBox [ 0 0 612 792 ] /Annots [ 99 0 R 100 0 R 101 0 R 102 0 R 103 0 R 104 0 R 105 0 R 106 0 R ] /Type /Page >> endobj 6 0 obj << /Contents 107 0 R /Parent 1 0 R /Resources 108 0 R /MediaBox [ 0 0 612 792 ] /Annots [ 109 0 R 110 0 R 111 0 R 112 0 R 113 0 R 114 0 R 115 0 R 116 0 R 117 0 R 118 0 R 119 0 R 120 0 R 121 0 R 122 0 R 123 0 R 124 0 R 125 0 R 126 0 R 127 0 R ] /Type /Page >> endobj 7 0 obj << /Contents 128 0 R /Parent 1 0 R /Resources 129 0 R /MediaBox [ 0 0 612 792 ] /Annots [ 158 0 R 159 0 R 160 0 R 161 0 R 162 0 R 163 0 R 164 0 R 165 0 R 166 0 R ] /Type /Page >> endobj 8 0 obj << /Contents 167 0 R /Parent 1 0 R /Resources 168 0 R /Group 180 0 R /MediaBox [ 0 0 612 792 ] /Annots [ 194 0 R 195 0 R 196 0 R 197 0 R 198 0 R 199 0 R 200 0 R 201 0 R 202 0 R 203 0 R 204 0 R 205 0 R 206 0 R ] /Type /Page >> endobj 9 0 obj << /Contents 207 0 R /Parent 1 0 R /Resources 208 0 R /MediaBox [ 0 0 612 792 ] /Annots [ 209 0 R 210 0 R 211 0 R 212 0 R 213 0 R 214 0 R 215 0 R 216 0 R 217 0 R 218 0 R ] /Type /Page >> endobj 10 0 obj << /Contents 219 0 R /Parent 1 0 R /Resources 220 0 R /MediaBox [ 0 0 612 792 ] /Annots [ 225 0 R 226 0 R 227 0 R 228 0 R 229 0 R 230 0 R 231 0 R 232 0 R 233 0 R 234 0 R 235 0 R 236 0 R 237 0 R 238 0 R 239 0 R 240 0 R 241 0 R 242 0 R 243 0 R 244 0 R 245 0 R 246 0 R 247 0 R 248 0 R 249 0 R 250 0 R 251 0 R 252 0 R 253 0 R 254 0 R 255 0 R 256 0 R 257 0 R ] /Type /Page >> endobj 11 0 obj << /Contents 258 0 R /Parent 1 0 R /Resources 259 0 R /MediaBox [ 0 0 612 792 ] /Annots [ 264 0 R 265 0 R 266 0 R 267 0 R 268 0 R 269 0 R 270 0 R 271 0 R 272 0 R 273 0 R 274 0 R 275 0 R 276 0 R 277 0 R 278 0 R 279 0 R 280 0 R 281 0 R 282 0 R 283 0 R ] /Type /Page >> endobj 12 0 obj << /Contents 284 0 R /Parent 1 0 R /Resources 285 0 R /MediaBox [ 0 0 612 792 ] /Annots [ 286 0 R 287 0 R ] /Type /Page >> endobj 13 0 obj << /Contents 288 0 R /Parent 1 0 R /Type /Page /Resources 289 0 R /MediaBox [ 0 0 612 792 ] >> endobj 14 0 obj << /Contents 290 0 R /Parent 1 0 R /Type /Page /Resources 291 0 R /MediaBox [ 0 0 612 792 ] >> endobj 15 0 obj << /Length 3479 /Filter /FlateDecode >> stream xڽZY~ׯGC&)ǀm ;puX.w#9*OUdU /fί$ Y\?