1 2 3<!doctype html> 4<!-- 5 Minimal Mistakes Jekyll Theme 4.28.1 by Michael Rose 6 Copyright 2013-2026 Michael Rose - mademistakes.com | @mmistakes 7 Copyright 2024-2026 iBug - ibugone.com | @iBug 8 Free for personal and commercial use under the MIT license 9 https://github.com/mmistakes/minimal-mistakes/blob/master/LICENSE 10--> 11 12<html lang="en" class="no-js" dir="ltr"> 13 <head> 14 15<meta charset="utf-8"> 16 17 18<!-- begin _includes/seo.html --><title>Posts - CAVA Lab</title> 19<meta name="description" content="Clinical AI Value Alignment"> 20 21 22 23<meta property="og:type" content="website"> 24<meta property="og:locale" content="en_US"> 25<meta property="og:site_name" content="CAVA Lab"> 26<meta property="og:title" content="Posts"> 27<meta property="og:url" content="https://cavalab.github.io/posts/"> 28 29 30 <meta property="og:description" content="Clinical AI Value Alignment"> 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45<link rel="canonical" href="https://cavalab.github.io/posts/"> 46 47 48 49 50 51 52 53 54 55 56 57 58<!-- end _includes/seo.html --> 59 60 61<meta name="viewport" content="width=device-width, initial-scale=1.0"> 62
63<script> 64 document.documentElement.className = document.documentElement.className.replace(/\bno-js\b/g, '') + ' js '; 65 66</script>
66 67 68<!-- For all browsers --> 69<link rel="stylesheet" href="/assets/css/main.css"> 70<link rel="preload" href="https://cdn.jsdelivr.net/npm/@fortawesome/fontawesome-free@latest/css/all.min.css" as="style" onload="this.onload=null;this.rel='stylesheet'"> 71<noscript><link rel="stylesheet" href="https://cdn.jsdelivr.net/npm/@fortawesome/fontawesome-free@latest/css/all.min.css"></noscript> 72 73 74 75 <!-- Load KaTeX for math rendering --> 76<!-- https://varunagrawal.github.io/2018/03/27/latex-jekyll/ --> 77<link rel="stylesheet" href="https://cdn.jsdelivr.net/npm/[email protected]/dist/katex.min.css" integrity="sha384-Xi8rHCmBmhbuyyhbI88391ZKP2dmfnOl4rT9ZfRI7mLTdk1wblIUnrIq35nqwEvC" crossorigin="anonymous"> 78 79<!-- The loading of KaTeX is deferred to speed up page rendering -->
80<script defer src="https://cdn.jsdelivr.net/npm/[email protected]/dist/katex.min.js" integrity="sha384-X/XCfMm41VSsqRNQgDerQczD69XqmjOOOwYQvr/uuC+j4OPoNhVgjdGFwhvN02Ja" crossorigin="anonymous"></script>
80 81 82<!-- To automatically render math in text elements, include the auto-render extension: -->
83<script defer src="https://cdn.jsdelivr.net/npm/[email protected]/dist/contrib/auto-render.min.js" integrity="sha384-+XBljXPPiv+OzfbB3cVmLHf4hdUFHlWNZN5spNQ7rmHTXpd7WvJum6fIACpNNfIR" crossorigin="anonymous" 84 onload="renderMathInElement(document.body);"></script>
84 85 86<!-- insert favicons. use https://realfavicongenerator.net/ --> 87 88<link rel="apple-touch-icon" sizes="180x180" href="/assets/images/apple-touch-icon.png"> 89<link rel="icon" type="image/png" sizes="32x32" href="/assets/images/favicon-32x32.png"> 90<link rel="icon" type="image/png" sizes="16x16" href="/assets/images/favicon-16x16.png"> 91<link rel="manifest" href="/assets/images/site.webmanifest"> 92<link rel="shortcut icon" href="/assets/images/favicon.ico"> 93<meta name="msapplication-TileColor" content="#da532c"> 94<meta name="msapplication-config" content="/assets/images/browserconfig.xml"> 95<meta name="theme-color" content="#ffffff"> 96 97 </head> 98 99 <body class="layout--archive"> 100 101<nav class="skip-links" aria-label="Skip links"> 102 <ul> 103 <li><a href="#site-nav" class="screen-reader-shortcut">Skip to primary navigation</a></li> 104 <li><a href="#main" class="screen-reader-shortcut">Skip to content</a></li> 105 <li><a href="#footer" class="screen-reader-shortcut">Skip to footer</a></li> 106 </ul> 107</nav> 108 109 110 111 112<div class="masthead"> 113 <div class="masthead__inner-wrap"> 114 <div class="masthead__menu"> 115 <nav id="site-nav" class="greedy-nav" aria-label="Primary navigation"> 116 117 <a class="site-title" href="/"> 118 CAVA Lab 119 <span class="site-subtitle">Clinical AI Value Alignment</span> 120 </a> 121 <ul class="visible-links"><li class="masthead__menu-item"> 122 <a 123 href="/members" 124 125 126 >Members</a> 127 </li><li class="masthead__menu-item"> 128 <a 129 href="/research" 130 131 132 >Research</a> 133 </li><li class="masthead__menu-item"> 134 <a 135 href="/papers" 136 137 138 >Papers</a> 139 </li><li class="masthead__menu-item"> 140 <a 141 href="/posts" 142 143 144 >Posts</a> 145 </li><li class="masthead__menu-item"> 146 <a 147 href="/join" 148 149 150 >Join</a> 151 </li></ul> 152 153 <button class="greedy-nav__toggle hidden" type="button"> 154 <span class="visually-hidden">Toggle menu</span> 155 <div class="navicon"></div> 156 </button> 157 <ul class="hidden-links hidden"></ul> 158 </nav> 159 </div> 160 </div> 161</div> 162 163 164 <div class="initial-content"> 165 166 167 168 169 170 171<div id="main" role="main"> 172 173 174 <div class="sidebar sticky"> 175 176 177 178 179<div itemscope itemtype="https://schema.org/Person" class="h-card"> 180 181 182 <div class="author__avatar"> 183 <a href="https://cavalab.github.io/"> 184 <img src="/assets/images/space_bot3.jpg" alt="" itemprop="image" class="u-photo"> 185 </a> 186 </div> 187 188 189 <div class="author__content"> 190 <h3 class="author__name p-name" itemprop="name"> 191 <a class="u-url" rel="me" href="https://cavalab.github.io/" itemprop="url"></a> 192 </h3> 193 194 </div> 195 196 <div class="author__urls-wrapper"> 197 <button class="btn btn--inverse">Follow</button> 198 <ul class="author__urls social-icons"> 199 200 <li itemprop="homeLocation" itemscope itemtype="https://schema.org/Place"> 201 <i class="fas fa-fw fa-location-dot" aria-hidden="true"></i> <span itemprop="name" class="p-locality">Boston, MA</span> 202 </li> 203 204 205 206 207 208 <li><a href="mailto:[email protected]" rel="nofollow noopener noreferrer me"><i class="fas fa-fw fa-envelope" aria-hidden="true"></i><span class="label">Email</span></a></li> 209 210 211 212 <li><a href="https://github.com/cavalab" rel="nofollow noopener noreferrer me" itemprop="sameAs"><i class="fab fa-fw fa-github" aria-hidden="true"></i><span class="label">GitHub</span></a></li> 213 214 215 216 <li><a href="https://scholar.google.com/citations?user=iZB7inEAAAAJ&hl=en" rel="nofollow noopener noreferrer me" itemprop="sameAs"><i class="fas fa-graduation-cap" aria-hidden="true"></i><span class="label">Google Scholar</span></a></li> 217 218 219 220 <li><a href="https://connects.catalyst.harvard.edu/Profiles/display/Person/200560" rel="nofollow noopener noreferrer me" itemprop="sameAs"><i class="fas fa-project-diagram" aria-hidden="true"></i><span class="label">Harvard Catalyst</span></a></li> 221 222 223 224 <li><a href="http://chip.org" rel="nofollow noopener noreferrer me" itemprop="sameAs"><i class="fa fa-heartbeat" aria-hidden="true"></i><span class="label">CHIP</span></a></li> 225 226 227 228 <li><a href="https://www.childrenshospital.org/" rel="nofollow noopener noreferrer me" itemprop="sameAs"><i class="fa fa-medkit" aria-hidden="true"></i><span class="label">Boston Children's Hospital</span></a></li> 229 230 231 232 <li><a href="https://hms.harvard.edu/" rel="nofollow noopener noreferrer me" itemprop="sameAs"><i class="fa fa-university" aria-hidden="true"></i><span class="label">
232Harvard Medical School</span></a></li> 233 234 235 236 <li><a href="https://bsky.app/profile/lacava.bsky.social" rel="nofollow noopener noreferrer me" itemprop="sameAs"><i class="far fa-comment" aria-hidden="true"></i><span class="label">Bluesky</span></a></li> 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 <!-- 290 <li> 291 <a href="http://link-to-whatever-social-network.com/user/" itemprop="sameAs" rel="nofollow noopener noreferrer me"> 292 <i class="fas fa-fw" aria-hidden="true"></i> Custom Social Profile Link 293 </a> 294 </li> 295--> 296 </ul> 297 </div> 298</div> 299 300 301 302 </div> 303 304 305 306 <div class="archive"> 307 308 <h1 id="page-title" class="page__title">Posts</h1> 309 310 <div class="entries-listpic"> 311 312 313 314 315 316<div class="listpic__item"> 317 <article class="archive__item" itemscope="" itemtype="https://schema.org/CreativeWork"> 318 319 <div class="archive__item-teaser"> 320 <img src="/assets/images/admission_plot.png" alt="" /> 321 </div> 322 323 <div class="archive__item-title no_toc" itemprop="headline"> 324 <b> 325 326 <a href="/2025/10/24/ed-admissions.html" rel="permalink">What Drives Disparities in Emergency Admissions? 327</a> 328 329 </b> 330 </div> 331 332 333 334 <p class="page__meta"> 335 336 337 338 339 340 341 342 343 <span class="page__meta-readtime"> 344 <i class="far fa-clock" aria-hidden="true"></i> 345 346 7 minute read 347 348 </span> 349 350 </p> 351 352 353 <p class="archive__item-excerpt" itemprop="description">Understanding unwarranted variations in emergency care 354</p> 355 </article> 356</div> 357 358 359 360 361 362<div class="listpic__item"> 363 <article class="archive__item" itemscope="" itemtype="https://schema.org/CreativeWork"> 364 365 <div class="archive__item-teaser"> 366 <img src="/assets/images/ecg-survival-figure-1.png" alt="" /> 367 </div> 368 369 <div class="archive__item-title no_toc" itemprop="headline"> 370 <b> 371 372 <a href="/2025/02/14/ecg-survival.html" rel="permalink">Benchmarking mortality risk prediction from electrocardiograms 373</a> 374 375 </b> 376 </div> 377 378 379 380 <p class="page__meta"> 381 382 383 384 385 386 387 388 389 <span class="page__meta-readtime"> 390 <i class="far fa-clock" aria-hidden="true"></i> 391 392 4 minute read 393 394 </span> 395 396 </p> 397 398 399 <p class="archive__item-excerpt" itemprop="description">Taking a closer look at survival modeling with ECGs 400</p> 401 </article> 402</div> 403 404 405 406 407 408<div class="listpic__item"> 409 <article class="archive__item" itemscope="" itemtype="https://schema.org/CreativeWork"> 410 411 <div class="archive__item-teaser"> 412 <img src="/assets/images/inexact-simplification-equivalent-models.svg" alt="" /> 413 </div> 414 415 <div class="archive__item-title no_toc" itemprop="headline"> 416 <b> 417 418 <a href="/2024/04/14/inexact-simplification.html" rel="permalink">Simplifying symbolic regression models without simplification rules 419</a> 420 421 </b> 422 </div> 423 424 425 426 <p class="page__meta"> 427 428 429 430 431 432 433 434 435 <span class="page__meta-readtime"> 436 <i class="far fa-clock" aria-hidden="true"></i> 437 438 6 minute read 439 440 </span> 441 442 </p> 443 444 445 <p class="archive__item-excerpt" itemprop="description">Relaxing the definition of equivalent mathematical expressions to get simpler and more interpretable models 446</p> 447 </article> 448</div> 449 450 451 452 453 454<div class="listpic__item"> 455 <article class="archive__item" itemscope="" itemtype="https://schema.org/CreativeWork"> 456 457 <div class="archive__item-teaser"> 458 <img src="/assets/images/HUMIES-2023-model-interpretation.svg" alt="" /> 459 </div> 460 461 <div class="archive__item-title no_toc" itemprop="headline"> 462 <b> 463 464 <a href="/2023/09/02/humies.html" rel="permalink">Interpreting a glass-box clinical prediction model of resistant hypertension 465</a> 466 467 </b> 468 </div> 469 470 471 472 <p class="page__meta"> 473 474 475 476 477 478 479 480
481 <span class="page__meta-readtime"> 482 <i class="far fa-clock" aria-hidden="true"></i> 483 484 5 minute read 485 486 </span> 487 488 </p> 489 490 491 <p class="archive__item-excerpt" itemprop="description">About our recent HUMIES award-winning algorithm for clinical prediction models 492</p> 493 </article> 494</div> 495 496 497 498 499 500<div class="listpic__item"> 501 <article class="archive__item" itemscope="" itemtype="https://schema.org/CreativeWork"> 502 503 <div class="archive__item-teaser"> 504 <img src="/assets/images/venn_dems.jpg" alt="" /> 505 </div> 506 507 <div class="archive__item-title no_toc" itemprop="headline"> 508 <b> 509 510 <a href="/2023/05/05/intersectionality.html" rel="permalink">Translating Intersectionality to Fair Machine Learning 511</a> 512 513 </b> 514 </div> 515 516 517 518 <p class="page__meta"> 519 520 521 522 523 524 525 526 527 <span class="page__meta-readtime"> 528 <i class="far fa-clock" aria-hidden="true"></i> 529 530 1 minute read 531 532 </span> 533 534 </p> 535 536 537 <p class="archive__item-excerpt" itemprop="description">A new perspective on how this social theory relates to fair machine learning. 538</p> 539 </article> 540</div> 541 542 543 544 545 546<div class="listpic__item"> 547 <article class="archive__item" itemscope="" itemtype="https://schema.org/CreativeWork"> 548 549 <div class="archive__item-teaser"> 550 <img src="/assets/images/clique.png" alt="" /> 551 </div> 552 553 <div class="archive__item-title no_toc" itemprop="headline"> 554 <b> 555 556 <a href="/2022/08/23/lexicase-running-time.html" rel="permalink">Adventures in Lexicase: Running Time 557</a> 558 559 </b> 560 </div> 561 562 563 564 <p class="page__meta"> 565 566 567 568 569 570 571 572 573 <span class="page__meta-readtime"> 574 <i class="far fa-clock" aria-hidden="true"></i> 575 576 7 minute read 577 578 </span> 579 580 </p> 581 582 583 <p class="archive__item-excerpt" itemprop="description">We consistently observe lexicase selection running times that are much lower than its worst-case bound of \(O(NC)\). Why? 584</p> 585 </article> 586</div> 587 588</div> 589 590 </div> 591</div> 592 593 594 </div> 595 596 597 598 <div id="footer" class="page__footer"> 599 <footer> 600 <!-- Parse the Latex divs with Katex--> 601<!-- https://varunagrawal.github.io/2018/03/27/latex-jekyll/ -->
602<script type="text/javascript"> 603 $("script[type='math/tex']").replaceWith( 604 function(){ 605 var tex = $(this).text(); 606 return katex.renderToString(tex, {displayMode: false}); 607 }); 608 609 $("script[type='math/tex; mode=display']").replaceWith( 610 function(){ 611 var tex = $(this).text(); 612 return katex.renderToString(tex.replace(/%.*/g, ''), {displayMode: true}); 613 }); 614</script>
614 615 616 617 618 619<div class="page__footer-follow"> 620 <ul class="social-icons"> 621 622 623 624 625 626 <li><a href="mailto:[email protected]" rel="nofollow noopener noreferrer"><i class="fas fa-fw fa-envelope" aria-hidden="true"></i> Email</a></li> 627 628 629 630 <li><a href="https://github.com/cavalab" rel="nofollow noopener noreferrer"><i class="fab fa-fw fa-github" aria-hidden="true"></i> GitHub</a></li> 631 632 633 634 <li><a href="https://scholar.google.com/citations?user=iZB7inEAAAAJ&hl=en" rel="nofollow noopener noreferrer"><i class="fas fa-graduation-cap" aria-hidden="true"></i> Google Scholar</a></li> 635 636 637 638 <li><a href="https://connects.catalyst.harvard.edu/Profiles/display/Person/200560" rel="nofollow noopener noreferrer"><i class="fas fa-project-diagram" aria-hidden="true"></i> Harvard Catalyst</a></li> 639 640 641 642 <li><a href="http://chip.org" rel="nofollow noopener noreferrer"><i class="fa fa-heartbeat" aria-hidden="true"></i> CHIP</a></li> 643 644 645 646 <li><a href="https://www.childrenshospital.org/" rel="nofollow noopener noreferrer"><i class="fa fa-medkit" aria-hidden="true"></i> Boston Children's Hospital</a></li> 647 648 649 650 <li><a href="https://hms.harvard.edu/" rel="nofollow noopener noreferrer"><i class="fa fa-university" aria-hidden="true"></i> Harvard Medical School</a></li> 651 652 653 654 <li><a href="https://bsky.app/profile/lacava.bsky.social" rel="nofollow noopener noreferrer"><i class="far fa-comment" aria-hidden="true"></i> Bluesky</a></li> 655 656 657 658 659 660 </ul> 661</div> 662 663 664<div class="page__footer-copyright">© 2026 <a href="https://cavalab.github.io">CAVA Lab</a>. Powered by <a href="https://jekyllrb.com" rel="nofollow">Jekyll</a> & <a href="https://mademistakes.com/work/jekyll-themes/minimal-mistakes/" rel="nofollow">Minimal Mistakes</a>.</div> 665 666 </footer> 667 </div> 668 669 670 671
671<script src="/assets/js/main.min.js"></script>
671 672 673 674 675 676 677 678 679 <!-- Global site tag (gtag.js) - Google Analytics -->
vendor: 64 bytes, lines 679-680
679 680<script async src="https://www.googletagmanager.com/gtag/js?id=
680G-8CVW0XBQDT
vendor: 12 bytes, line 680
680"></script>
681<script> 682
vendor: 134 bytes, lines 682-686
682window.dataLayer = window.dataLayer || []; 683 function gtag(){dataLayer.push(arguments);} 684 gtag('js', new Date()); 685 686 gtag('config', '
686G-8CVW0XBQDT
vendor: 29 bytes, line 686
686', { 'anonymize_ip': true});
687</script>
687 688 689 690 691 692 693 694 695 696 697 </body> 698</html>
Line numbers count LF bytes from the start of the resource, as the search results do. Vendor segments are library code the classifier recognised; they are stored but not indexed. Bytes are shown as Latin1 characters, one per byte.