28:
264:. To design a Wiener filter, one must know the statistics of the noiseless signal that one wants to recover. However, if the statistics of the noiseless signal is unknown, this cannot be designed. Widrow thus designed an adaptive filter that uses gradient descent to minimize the mean square error. He also attended the
324:
Unable to train multilayered neural networks, Widrow turned to adaptive filtering and adaptive signal processing, using techniques based on the LMS filter for applications such as adaptive antenna, adaptive noise canceling, and applications to medicine.
303:
machine should not be connected randomly to the A-units. Instead, the S-units should be removed, so that the photocell inputs would be directly inputted into the A-units. Rosenblatt objected that "the human retina is built that way".
311:
Rule I (1962), which had two weight layers. The first was trainable, but the second was fixed. Widrow stated their problem would have been solved by the backpropagation algorithm. "This was long before
234:
computer. The experience of building magnetic core memory shaped his understanding of computers into a "memory's eye view", that is, he "look for the memory and see what you have to connect around it".
963:
226:
He entered MIT in 1947, studied electrical engineering and electronics, and graduated in 1951. After that, he got a research assistantship in the MIT Digital
Computer Laboratory, in the
973:
864:"The 2001 Benjamin Franklin Medal in Engineering presented to Bernard Widrow - Journal of the Franklin Institute - Tom 339, Numer 3 (2002) - Biblioteka Nauki - Yadda"
332:, he noticed that neural network research was returning, and he also learned of the backpropagation algorithm. After that, he returned to neural network research.
925:
551:"(1960) Bernard Widrow and Marcian E. Hoff, "Adaptive switching circuits," [i]1960 IRE WESCON Convention Record[/i], New York: IRE, pp. 96-104."
978:
246:
of the sensing signal of magnetic core memory. Back then, the hysteresis loops for magnetic core memory was not square enough, making sensing signal noisy.
307:
Despite many attempts, they never succeeded in developing a training algorithm for a multilayered neural network. The furthest they got was with
528:
948:
788:
Widrow, B.; Glover, J.R.; McCool, J.M.; Kaunitz, J.; Williams, C.S.; Hearn, R.H.; Zeidler, J.R.; Eugene Dong, Jr.; Goodlin, R.C. (1975).
863:
953:
289:
635:
196:
He is the namesake of "Uncle Bernie's Rule": the training sample size should be 10 times the number of weights in a network.
69:
968:
617:
280:. They improved the previous adaptive filter so that it makes a gradient descent for each datapoint, resulting in the
570:
907:
392:
250:
592:"Bernard Widrow, an oral history conducted in 1997 by Andrew Goldstein, IEEE History Center, Piscataway, NJ, USA"
402:
414:
652:"Collection: Massachusetts Institute of Technology, Digital Computer Laboratory records | MIT ArchivesSpace"
341:
1965 "A critical comparison of two kinds of adaptive classification networks", K. Steinbuch and B. Widrow,
364:
Quantization Noise: Roundoff Error in
Digital Computation, Signal Processing, Control, and Communications.
178:
163:
216:. While young, he was interested in electronics. During WWII, he found an entry on "Radios" in the
827:
Yelderman, Mark; Widrow, Bernard; Cioffi, John M.; Hesler, Edward; Leddy, Jeffrey A. (July 1983).
424:
He was one of the Board of
Governors of the International Neural Network Society (INNIS) in 2003.
218:
958:
470:
155:
86:
651:
386:
243:
943:
227:
213:
8:
921:
447:. Information Systems Laboratory, Electrical Engineering Department, Stanford University.
159:
96:
288:. To avoid having to hand-tune the weights in ADALINE, they invented the memistor, with
230:
group. The DCL was a division of the
Servomechanisms Laboratory, which was building the
724:
265:
249:
For his PhD (1956, advised by
William Linvill), he worked on the statistical theory of
190:
879:
844:
809:
770:
706:
631:
627:
566:
562:
186:
550:
875:
836:
801:
762:
696:
688:
623:
558:
509:
296:
103:
498:"30 years of adaptive neural networks: perceptron, Madaline, and backpropagation"
239:
182:
131:
108:
828:
750:
676:
900:
789:
591:
529:"Generalization and Parameter Estimation in Feedforward Nets: Some Experiments"
329:
136:
497:
27:
937:
917:
848:
840:
813:
774:
710:
692:
677:"A Study of Rough Amplitude Quantization by Means of Nyquist Sampling Theory"
257:
121:
59:
805:
766:
277:
261:
167:
126:
829:"ECG Enhancement by Adaptive Cancellation of Electrosurgical Interference"
185:
technique. He made other fundamental contributions to the development of
175:
313:
231:
701:
444:
300:
281:
292:(ADALINE weights) being the thickness of the copper on the graphite.
513:
308:
285:
171:
439:
437:
749:
Widrow, B.; Mantey, P.E.; Griffiths, L.J.; Goode, B.B. (1967).
420:
International Neural
Network Society (INNIS) Board member 2004
434:
964:
Members of the United States
National Academy of Engineering
826:
377:
253:, inspired by work by William Linvill and David Middleton.
748:
787:
790:"Adaptive noise cancelling: Principles and applications"
166:(LMS) adaptive algorithm with his then doctoral student
32:
Widrow demonstrating the "Knobby
Adaline" device (1963)
974:
Benjamin
Franklin Medal (Franklin Institute) laureates
616:
Anderson, James A.; Rosenfeld, Edward, eds. (2000).
189:
in the fields of geophysics, adaptive antennas, and
615:
533:Advances in Neural Information Processing Systems
935:
619:Talking Nets: An Oral History of Neural Networks
468:
154:(born December 24, 1929) is a U.S. professor of
557:, The MIT Press, pp. 123–134, 1988-04-07,
526:
979:Massachusetts Institute of Technology alumni
276:In 1959, he got his first graduate student,
162:. He is the co-inventor of the Widrow–Hoff
833:IEEE Transactions on Biomedical Engineering
207:
495:
464:
462:
460:
458:
456:
454:
408:IEEE Signal Processing Society Award, 1999
26:
700:
496:Widrow, B.; Lehr, M.A. (September 1990).
343:IEEE Transactions on Electronic Computers
238:For his masters thesis (1953, advised by
398:IEEE Neural Networks Pioneer Medal, 1991
316:. Backprop to me is almost miraculous."
299:, Widrow argued that the S-units in the
268:in 1956 and was inspired to work on AI.
451:
936:
674:
359:New Jersey: Prentice-Hall, Inc., 1994.
352:New Jersey: Prentice-Hall, Inc., 1985.
861:
725:"Oral-History:David Middleton (2000)"
319:
70:Massachusetts Institute of Technology
949:Artificial intelligence researchers
13:
681:IRE Transactions on Circuit Theory
348:1985 B. Widrow and S. D. Stearns.
14:
990:
868:Journal of the Franklin Institute
527:Morgan, N.; Bourlard, H. (1989).
366:Cambridge University Press, 2008.
908:IEEE Alexander Graham Bell Medal
393:IEEE Alexander Graham Bell Medal
170:. The LMS algorithm led to the
954:IEEE Centennial Medal laureates
855:
820:
781:
742:
717:
403:National Academy of Engineering
362:2008 B. Widrow and I. Kollar.
335:
668:
644:
628:10.7551/mitpress/6626.003.0004
609:
584:
563:10.7551/mitpress/4943.003.0012
543:
520:
489:
355:1994 B. Widrow and E. Walach.
1:
880:10.1016/S0016-0032(01)00044-8
471:"Bernard Widrow Oral History"
427:
271:
445:"Widrow's Stanford web page"
242:), he worked on raising the
223:and built a one-tube radio.
199:
193:. A summary of his work is.
7:
969:Stanford University faculty
475:IEEE Global History Network
411:IEEE Millennium Medal, 2000
350:Adaptive Signal Processing.
256:During PhD, he learned the
10:
995:
751:"Adaptive antenna systems"
204:This section is based on.
179:artificial neural networks
914:
905:
897:
892:
469:Andrew Goldstein (1997).
383:Elected Fellow AAAS, 1980
370:
357:Adaptive Inverse Control.
164:least mean squares filter
145:
114:
102:
92:
82:
75:
65:
55:
37:
25:
18:
841:10.1109/TBME.1983.325039
693:10.1109/TCT.1956.1086334
555:Neurocomputing, Volume 1
328:At a 1985 conference in
208:Early life and education
862:Abend, Kenneth (2002).
835:. BME-30 (7): 392–398.
806:10.1109/PROC.1975.10036
794:Proceedings of the IEEE
755:Proceedings of the IEEE
502:Proceedings of the IEEE
415:Benjamin Franklin Medal
219:World Book Encyclopedia
767:10.1109/PROC.1967.6092
295:During a meeting with
156:electrical engineering
87:Electrical engineering
926:William C. Jakes, Jr.
656:archivesspace.mit.edu
387:IEEE Centennial Medal
244:signal-to-noise ratio
228:magnetic core memory
214:Norwich, Connecticut
922:Richard H. Frenkiel
675:Widrow, B. (1956).
345:, pp. 737–740.
160:Stanford University
97:Stanford University
539:. Morgan-Kaufmann.
401:Inducted into the
320:Adaptive filtering
266:Dartmouth workshop
251:quantization noise
191:adaptive filtering
932:
931:
915:Succeeded by
800:(12): 1692–1716.
761:(12): 2143–2159.
637:978-0-262-26715-1
622:. The MIT Press.
187:signal processing
149:
148:
115:Doctoral students
77:Scientific career
48:December 24, 1929
986:
898:Preceded by
890:
889:
884:
883:
874:(339): 283–294.
859:
853:
852:
824:
818:
817:
785:
779:
778:
746:
740:
739:
737:
736:
721:
715:
714:
704:
672:
666:
665:
663:
662:
648:
642:
641:
613:
607:
606:
604:
603:
588:
582:
581:
580:
579:
547:
541:
540:
524:
518:
517:
508:(9): 1415–1442.
493:
487:
486:
484:
482:
466:
449:
448:
441:
297:Frank Rosenblatt
104:Doctoral advisor
51:
47:
45:
30:
16:
15:
994:
993:
989:
988:
987:
985:
984:
983:
934:
933:
928:
911:
903:
888:
887:
860:
856:
825:
821:
786:
782:
747:
743:
734:
732:
723:
722:
718:
673:
669:
660:
658:
650:
649:
645:
638:
614:
610:
601:
599:
590:
589:
585:
577:
575:
573:
549:
548:
544:
525:
521:
514:10.1109/5.58323
494:
490:
480:
478:
467:
452:
443:
442:
435:
430:
376:Elected Fellow
373:
338:
322:
274:
240:William Linvill
212:He was born in
210:
202:
183:backpropagation
141:
132:Richard Mattson
109:William Linvill
66:Alma mater
49:
43:
41:
33:
21:
12:
11:
5:
992:
982:
981:
976:
971:
966:
961:
956:
951:
946:
930:
929:
916:
913:
904:
901:Charles K. Kao
899:
895:
894:
886:
885:
854:
819:
780:
741:
716:
687:(4): 266–276.
667:
643:
636:
608:
583:
571:
542:
519:
488:
450:
432:
431:
429:
426:
422:
421:
418:
412:
409:
406:
399:
396:
390:
384:
381:
372:
369:
368:
367:
360:
353:
346:
337:
334:
330:Snowbird, Utah
321:
318:
273:
270:
209:
206:
201:
198:
152:Bernard Widrow
147:
146:
143:
142:
140:
139:
137:Alvy Ray Smith
134:
129:
124:
118:
116:
112:
111:
106:
100:
99:
94:
90:
89:
84:
80:
79:
73:
72:
67:
63:
62:
57:
53:
52:
39:
35:
34:
31:
23:
22:
20:Bernard Widrow
19:
9:
6:
4:
3:
2:
991:
980:
977:
975:
972:
970:
967:
965:
962:
960:
959:Living people
957:
955:
952:
950:
947:
945:
942:
941:
939:
927:
923:
919:
918:Joel S. Engel
910:
909:
902:
896:
891:
881:
877:
873:
869:
865:
858:
850:
846:
842:
838:
834:
830:
823:
815:
811:
807:
803:
799:
795:
791:
784:
776:
772:
768:
764:
760:
756:
752:
745:
730:
726:
720:
712:
708:
703:
698:
694:
690:
686:
682:
678:
671:
657:
653:
647:
639:
633:
629:
625:
621:
620:
612:
597:
593:
587:
574:
572:9780262267137
568:
564:
560:
556:
552:
546:
538:
534:
530:
523:
515:
511:
507:
503:
499:
492:
476:
472:
465:
463:
461:
459:
457:
455:
446:
440:
438:
433:
425:
419:
416:
413:
410:
407:
404:
400:
397:
394:
391:
388:
385:
382:
379:
375:
374:
365:
361:
358:
354:
351:
347:
344:
340:
339:
333:
331:
326:
317:
315:
310:
305:
302:
298:
293:
291:
287:
283:
279:
269:
267:
263:
259:
258:Wiener filter
254:
252:
247:
245:
241:
236:
233:
229:
224:
222:
220:
215:
205:
197:
194:
192:
188:
184:
180:
177:
173:
169:
165:
161:
157:
153:
144:
138:
135:
133:
130:
128:
125:
123:
122:Parham Aarabi
120:
119:
117:
113:
110:
107:
105:
101:
98:
95:
91:
88:
85:
81:
78:
74:
71:
68:
64:
61:
58:
54:
50:(age 94)
40:
36:
29:
24:
17:
906:
871:
867:
857:
832:
822:
797:
793:
783:
758:
754:
744:
733:. Retrieved
731:. 2021-01-26
728:
719:
702:1721.1/12139
684:
680:
670:
659:. Retrieved
655:
646:
618:
611:
600:. Retrieved
595:
586:
576:, retrieved
554:
545:
536:
532:
522:
505:
501:
491:
479:. Retrieved
474:
423:
363:
356:
349:
342:
336:Publications
327:
323:
306:
294:
275:
262:Lee Yuk-wing
255:
248:
237:
225:
217:
211:
203:
195:
151:
150:
127:Marcian Hoff
93:Institutions
76:
944:1929 births
314:Paul Werbos
290:conductance
232:Whirlwind I
181:and to the
56:Nationality
938:Categories
735:2023-11-03
661:2023-11-03
602:2023-11-03
578:2023-11-03
428:References
301:perceptron
282:delta rule
272:Work on AI
44:1929-12-24
849:0018-9294
814:0018-9219
775:0018-9219
711:0096-2007
481:22 August
200:Biography
309:Madaline
284:and the
278:Ted Hoff
176:MADALINE
168:Ted Hoff
60:American
893:Awards
286:ADALINE
172:ADALINE
847:
812:
773:
709:
634:
598:. 1997
569:
477:. IEEE
417:, 2001
405:, 1995
395:, 1986
389:, 1984
380:, 1976
371:Honors
83:Fields
912:1986
260:from
924:and
845:ISSN
810:ISSN
771:ISSN
729:ETHW
707:ISSN
632:ISBN
596:ETHW
567:ISBN
483:2011
378:IEEE
174:and
38:Born
876:doi
837:doi
802:doi
763:doi
697:hdl
689:doi
624:doi
559:doi
510:doi
158:at
940::
920:,
870:.
866:.
843:.
831:.
808:.
798:63
796:.
792:.
769:.
759:55
757:.
753:.
727:.
705:.
695:.
683:.
679:.
654:.
630:.
594:.
565:,
553:,
535:.
531:.
506:78
504:.
500:.
473:.
453:^
436:^
46:)
882:.
878::
872:3
851:.
839::
816:.
804::
777:.
765::
738:.
713:.
699::
691::
685:3
664:.
640:.
626::
605:.
561::
537:2
516:.
512::
485:.
221:,
42:(
Text is available under the Creative Commons Attribution-ShareAlike License. Additional terms may apply.
