The Discriminative Lexicon: A Unified Computational Model for the Lexicon and Lexical Processing in Comprehension and Production Grounded Not in (De)Composition but in Linear Discriminative Learning

<div>Similarity rating in the MEN dataset as a function of the correlation <svg height="6.1673pt" id="M138" style="vertical-align:-0.2063904pt" version="1.1" viewbox="-0.0498162 -5.96091 5.60619 6.1673" width="5.60619pt" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink"><g transform="matrix(.013,0,0,-0.013,0,0)"><path d="M393 379C402 394 400 411 393 422C384 437 365 448 348 448C301 448 237 372 186 285H182L193 335C210 408 205 448 178 448C150 448 80 402 29 344L45 321C80 355 114 373 122 373C128 373 130 365 124 330C106 228 76 98 50 -5L57 -12C82 -5 112 3 132 6L172 203C196 256 234 304 254 329C275 355 293 367 306 367C318 367 330 360 342 348C347 343 355 343 365 350S386 367 393 379Z"></path></g></svg> between the row vectors in <svg height="8.97279pt" id="M139" style="vertical-align:-0.2324896pt" version="1.1" viewbox="-0.0498162 -8.7403 6.68926 8.97279" width="6.68926pt" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink"><g transform="matrix(.013,0,0,-0.013,0,0)"><path d="M442 495C434 557 427 602 423 639C405 645 348 665 284 665C142 665 55 579 55 469C55 373 127 319 229 267C304 227 348 190 348 131C348 67 309 25 248 25C152 25 93 138 73 202L38 190C45 130 58 46 68 21C81 15 100 7 128 -1C158 -10 193 -15 229 -15C375 -15 472 74 472 188C472 287 400 343 295 394C207 439 171 470 171 525C171 583 212 624 268 624C349 624 387 550 407 487L442 495Z"></path></g></svg> of the words in the test pairs.</div>

Complexity

fig1

Figure 1

Figure 1: The Discriminative Lexicon: A Unified Computational Model for the Lexicon and Lexical Processing in Comprehension and Production Grounded Not in (De)Composition but in Linear Discriminative Learning 

Figure 1 | The Discriminative Lexicon: A Unified Computational Model for the Lexicon and Lexical Processing in Comprehension and Production Grounded Not in (De)Composition but in Linear Discriminative Learning 