476:. An earlier CNN on GPU by Chellapilla et al. (2006) was 4 times faster than an equivalent implementation on CPU. The deep CNN of Dan Ciresan et al. (2011) at IDSIA was already 60 times faster and achieved the first superhuman performance in a computer vision contest in August 2011. Between 15 May 2011 and 10 September 2012, these CNNs won four more image competitions and improved the state of the art on multiple image benchmarks. The approach has become central to the field of
533:, Yann LeCun wrote that "Jürgen is manically obsessed with recognition and keeps claiming credit he doesn't deserve for many, many things... It causes him to systematically stand up at the end of every talk and claim credit for what was just presented, generally not in a justified manner." Schmidhuber replied that LeCun did this "without any justification, without providing a single example," and published details of numerous priority disputes with Hinton, Bengio and LeCun.
44:
573:
perform their own research, and explore the universe. He has worked on both types for decades, He expects the next stage of evolution to be self-improving AIs that will succeed human civilization as the next stage in the universal increase towards ever-increasing complexity, and he expects AI to colonize the visible universe.
373:. The name LSTM was introduced in a tech report (1995) leading to the most cited LSTM publication (1997), co-authored by Hochreiter and Schmidhuber. It was not yet the standard LSTM architecture which is used in almost all current applications. The standard LSTM architecture was introduced in 2000 by
536:
The term "schmidhubered" has been jokingly used in the AI community to describe
Schmidhuber's habit of publicly challenging the originality of other researchers' work, a practice seen by some in the AI community as a "rite of passage" for young researchers. Some suggest that Schmidhuber's significant
572:
Since the 1970s, Schmidhuber wanted to create "intelligent machines that could learn and improve on their own and become smarter than him within his lifetime." He differentiates between two types of AIs: tool AI, such as those for improving healthcare, and autonomous AIs that set their own goals,
1347:
Wu, Yonghui; Schuster, Mike; Chen, Zhifeng; Le, Quoc V.; Norouzi, Mohammad; Macherey, Wolfgang; Krikun, Maxim; Cao, Yuan; Gao, Qin; Macherey, Klaus; Klingner, Jeff; Shah, Apurva; Johnson, Melvin; Liu, Xiaobing; Kaiser, Łukasz; Gouws, Stephan; Kato, Yoshikiyo; Kudo, Taku; Kazawa, Hideto; Stevens,
266:
are advisers to the company. Sales were under US$ 11 million in 2016; however, Schmidhuber states that the current emphasis is on research and not revenue. Nnaisense raised its first round of capital funding in
January 2017. Schmidhuber's overall goal is to create an
456:
of self-generated activation patterns, and the fast weights network itself operates over inputs. Schmidhuber used the terminology "learning internal spotlights of attention" in 1993. Recently he renamed it to "linearized
Transformer" and claims it was a precursor to
568:
Schmidhuber is a proponent of open source AI, and believes that they will become competitive against commercial closed-source AI. He does not believe AI poses a new existential threat, and is less threatening than nuclear weapons.
173:. He is also director of the Artificial Intelligence Initiative and professor of the Computer Science program in the Computer, Electrical, and Mathematical Sciences and Engineering (CEMSE) division at the
528:
for their work in deep learning. He wrote a "scathing" 2015 article arguing that Hinton, Bengio and Lecun "heavily cite each other" but "fail to credit the pioneers of the field". In a statement to the
389:
training algorithm in 2006. CTC was applied to end-to-end speech recognition with LSTM. By the 2010s, the LSTM became the dominant technique for a variety of natural language processing tasks including
2516:
2506:
1322:
Graves, Alex; Fernández, Santiago; Gomez, Faustino; Schmidhuber, Juergen (2006). "Connectionist temporal classification: Labelling unsegmented sequence data with recurrent neural networks".
244:
166:
120:
350:
where the environmental reaction is 1 or 0 depending on whether the first network's output is in a given set. GANs were the state of the art in generative modeling during 2015-2020 period.
560:
the "father of deep learning," and gives credit to many even earlier AI pioneers. Though
Ivakhnenko himself credited Rosenblatt's perceptron as an example of multilayered neural network.
2269:"iHuman- AI & Ethics of Cinema (2020 Hot Docs Film Festival). Quote: The documentary interviews range AI top researchers and thinkers as Jürgen Schmidhuber - Father of Modern AI..."
2139:
798:
2248:
727:
Schmidhuber, Jürgen (2020). "Generative
Adversarial Networks are Special Cases of Artificial Curiosity (1990) and also Closely Related to Predictability Minimization (1991)".
2521:
2491:
2325:
1403:
174:
1808:
2402:
2065:
2511:
909:
250:
In 2014, Schmidhuber formed a company, Nnaisense, to work on commercial applications of artificial intelligence in fields such as finance, heavy industry and
1457:
He, Kaiming; Zhang, Xiangyu; Ren, Shaoqing; Sun, Jian (2016). "Delving Deep into
Rectifiers: Surpassing Human-Level Performance on ImageNet Classification".
2169:
1083:
2170:"Artificial general intelligence: Are we close, and does it even make sense to try? Quote: Jürgen Schmidhuber—sometimes called "the father of modern AI..."
1348:
Keith; Kurian, George; Patil, Nishant; Wang, Wei; Young, Cliff; Smith, Jason; Riesa, Jason; Rudnick, Alex; Vinyals, Oriol; Corrado, Greg; Hughes, Macduff;
386:
926:
458:
414:
2376:
627:
556:
He has been referred to as the "father of modern AI" or similar, and also the "father of deep learning." Schmidhuber himself, however, has called
546:
508:
Schmidhuber has controversially argued that he and other researchers have been denied adequate recognition for their contribution to the field of
1956:
1889:
Fukushima, Neocognitron (1980). "A self-organizing neural network model for a mechanism of pattern recognition unaffected by shift in position".
413:
In 2014, the state of the art was training “very deep neural network” with 20 to 30 layers. Stacking too many layers led to a steep reduction in
2496:
1844:
1580:
550:
346:
to predict the reactions of the environment to these patterns. This was called "artificial curiosity." In 2014, this principle was used in a
233:
2351:
545:
Schmidhuber received the
Helmholtz Award of the International Neural Network Society in 2013, and the Neural Networks Pioneer Award of the
2011:
1985:
2224:"Letting loose the AI demon. Quote: But this man is no crackpot: He is the father of modern AI and deep learning – foremost in his field"
2066:"Scientific Integrity and the History of Deep Learning: The 2021 Turing Lecture, and the 2018 Turing Award. Technical Report IDSIA-77-21"
646:
2501:
2136:
2038:
1208:
Graves, A.; Schmidhuber, J. (2005). "Framewise phoneme classification with bidirectional LSTM and other neural network architectures".
2092:
399:
2296:
1261:
Klaus Greff; Rupesh Kumar
Srivastava; Jan Koutník; Bas R. Steunebrink; Jürgen Schmidhuber (2015). "LSTM: A Search Space Odyssey".
382:
197:
93:
1411:
268:
307:
at multiple self-organizing time scales. This can substantially facilitate downstream deep learning. The RNN hierarchy can be
347:
217:
205:
201:
75:
2039:"How 3 Turing Awardees Republished Key Methods and Ideas Whose Creators They Failed to Credit. Technical Report IDSIA-23-23"
1374:
1803:
Ciresan, Dan; Meier, Ueli; Schmidhuber, Jürgen (June 2012). "Multi-column deep neural networks for image classification".
557:
709:
Schmidhuber, Jürgen (1991). "A possibility for implementing curiosity and boredom in model-building neural controllers".
2377:
Der unbequeme Vater der künstlichen
Intelligenz lebt in der Schweiz (The inconvenient father of AI lives in Switzerland)
2069:
1388:
1352:(8 October 2016). "Google's Neural Machine Translation System: Bridging the Gap between Human and Machine Translation".
1155:
1110:
497:
469:
378:
1599:(1993). "Reducing the ratio between learning complexity and number of time-varying variables in fully recurrent nets".
866:
Schlag, Imanol; Irie, Kazuki; Schmidhuber, Jürgen (2021). "Linear
Transformers Are Secretly Fast Weight Programmers".
189:
2268:
826:
Schmidhuber, Jürgen (1 November 1992). "Learning to control fast-weight memories: an alternative to recurrent nets".
232:. He taught there from 2004 until 2009. From 2009, until 2021, he was a professor of artificial intelligence at the
445:
426:
362:
162:
417:
accuracy, known as the "degradation" problem. In 2015, Rupesh Kumar Srivastava, Klaus Greff, and Schmidhuber used
1080:
288:
1720:
1615:
441:
370:
339:
304:
296:
292:
2012:"LeCun's 2022 paper on autonomous machine intelligence rehashes but does not cite essential work of 1990-2015"
1153:
Felix A. Gers; Jürgen Schmidhuber; Fred Cummins (2000). "Learning to Forget: Continual Prediction with LSTM".
1932:
Weng, J; Ahuja, N; Huang, TS (1993). "Learning recognition and segmentation of 3-D objects from 2-D images".
959:"Learning complex, extended sequences using the principle of history compression (based on TR FKI-148, 1991)"
357:
which he considered "one of the most important documents in the history of machine learning". It studied the
1668:
Proceedings of the Twenty-Second International Joint Conference on Artificial Intelligence-Volume Volume Two
430:
158:
110:
2486:
1822:
1759:
1327:
1217:
1164:
366:
312:
193:
185:
85:
1045:
Schmidhuber, Jürgen (2010). "Formal Theory of Creativity, Fun, and Intrinsic Motivation (1990-2010)".
2526:
1280:
468:
In 2011, Schmidhuber's team at IDSIA with his postdoc Dan Ciresan also achieved dramatic speedups of
2223:
1827:
1693:
1222:
1764:
1332:
1169:
681:
549:
in 2016 for "pioneering contributions to deep learning and neural networks." He is a member of the
395:
2249:
Jürgen Schmidhuber, el hombre al que Alexa y Siri llamarían ‘papá’ si él quisiera hablar con ellas
2194:
1986:"Jürgen Schmidhuber on the robot future: 'They will pay as much attention to us as we do to ants'"
799:"Juergen Schmidhuber, Renowned 'Father Of Modern AI,' Says His Life's Work Won't Lead To Dystopia"
1914:
1858:
1812:
1785:
1558:
1504:
1483:
1458:
1438:
1353:
1304:
1270:
1243:
1190:
1135:
1062:
1005:
981:
843:
762:
736:
489:
391:
154:
2450:
2428:
2155:
2015:
1525:
2100:
1906:
1850:
1840:
1777:
1576:
1503:
Srivastava, Rupesh Kumar; Greff, Klaus; Schmidhuber, Jürgen (2 May 2015). "Highway Networks".
1296:
1235:
1182:
1127:
754:
654:
300:
2326:"Rise of artificial intelligence is inevitable but should not be feared, 'father of AI' says"
2042:
881:
192:
tasks in research and commercial applications in the 2010s. He also introduced principles of
2462:
1898:
1832:
1769:
1568:
1288:
1227:
1174:
1119:
1054:
973:
835:
746:
449:
343:
335:
251:
89:
188:(LSTM), a type of neural network architecture which was the dominant technique for various
2143:
1659:
Ciresan, Dan; Ueli Meier; Jonathan Masci; Luca M. Gambardella; Jurgen Schmidhuber (2011).
1087:
513:
485:
477:
462:
422:
354:
280:
255:
225:
1875:
Y. LeCun, B. Boser, J. S. Denker, D. Henderson, R. E. Howard, W. Hubbard, L. D. Jackel,
1284:
1092:
Diploma thesis. Institut f. Informatik, Technische Univ. Munich. Advisor: J. Schmidhuber
1026:
43:
2403:"Father of deep learning AI on General purpose AI and AI to conquer space in the 2050s"
633:
327:
229:
1876:
2480:
2466:
1918:
958:
910:
The Present and Future of AI and Deep Learning Featuring Professor Jürgen Schmidhuber
766:
517:
509:
453:
331:
284:
263:
1747:
1194:
1105:
985:
847:
1952:
1862:
1789:
1661:"Flexible, High Performance Convolutional Neural Networks for Image Classification"
1308:
1247:
1139:
537:
accomplishments have been underappreciated due to his confrontational personality.
525:
493:
403:
323:
network. In 1993, a chunker solved a deep learning task whose depth exceeded 1000.
259:
178:
64:
17:
2146:. Piscataway, NJ: IEEE Computational Intelligence Society. Accessed January 2019.]
1066:
1231:
1004:
Schmidhuber, Juergen (2022). "Annotated History of Modern AI and Deep Learning".
750:
1773:
170:
1641:
1548:
1292:
1178:
1123:
429:
with hundreds of layers, much deeper than previous networks. Concurrently, the
334:, where one network's gain is the other network's loss. The first network is a
1836:
1058:
977:
839:
521:
481:
374:
2124:
2104:
1854:
1660:
1324:
In Proceedings of the International Conference on Machine Learning, ICML 2006
216:
Schmidhuber completed his undergraduate (1987) and PhD (1991) studies at the
1349:
1781:
1300:
1239:
1186:
758:
1910:
1478:
He, Kaiming; Zhang, Xiangyu; Ren, Shaoqing; Sun, Jian (10 December 2015).
1131:
604:
1572:
1642:"High Performance Convolutional Neural Networks for Document Processing"
2253:
1902:
291:. To overcome this problem, Schmidhuber (1991) proposed a hierarchy of
1555:
2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)
2429:"Critique of Paper by "Deep Learning Conspiracy". (Nature 521 p 436)"
237:
221:
60:
2451:"Heuristic self-organization in problems of engineering cybernetics"
1646:
Tenth International Workshop on Frontiers in Handwriting Recognition
433:(ResNet) developed, which is a special case of the highway network.
1563:
1509:
1488:
1463:
1358:
1275:
1010:
741:
135:
1817:
1443:
1435:
Very Deep Convolutional Networks for Large-Scale Image Recognition
271:
by training a single AI in sequence on a variety of narrow tasks.
1524:
Srivastava, Rupesh K; Greff, Klaus; Schmidhuber, Juergen (2015).
927:"AI Pioneer Wants to Build the Renaissance Machine of the Future"
2297:"'Father of AI' says tech fears misplaced: 'You cannot stop it'"
2158:. European Academy of Sciences and Arts. Accessed December 2016.
2093:"Jürgen Schmidhuber: Tessiner Vater der künstlichen Intelligenz"
667:
418:
407:
398:, and was widely implemented in commercial technologies such as
2127:. International Neural Network Society. Accessed December 2016.
1805:
2012 IEEE Conference on Computer Vision and Pattern Recognition
184:
He is best known for his foundational and highly-cited work on
2381:
1547:
He, Kaiming; Zhang, Xiangyu; Ren, Shaoqing; Sun, Jian (2016).
473:
452:
to control the fast weights of another neural network through
377:, Schmidhuber, and Fred Cummins. Today's "vanilla LSTM" using
353:
Schmidhuber supervised the 1991 diploma thesis of his student
2352:"The 'father of A.I' urges humans not to fear the technology"
1721:"History of computer vision contests won by deep CNNs on GPU"
48:
Schmidhuber speaking at the AI for GOOD Global Summit in 2017
1957:"This Man Is the Godfather the AI Community Wants to Forget"
1877:
Backpropagation Applied to Handwritten Zip Code Recognition
245:
Dalle Molle Institute for Artificial Intelligence Research
167:
Dalle Molle Institute for Artificial Intelligence Research
121:
Dalle Molle Institute for Artificial Intelligence Research
1263:
IEEE Transactions on Neural Networks and Learning Systems
480:. It is based on CNN designs introduced much earlier by
1404:"The iBrain Is Here—and It's Already Inside Your Phone"
1375:"The neural networks behind Google Voice transcription"
628:
When A.I. Matures, It May Call Jürgen Schmidhuber ‘Dad’
2517:
Academic staff of the Technical University of Munich
2507:
Members of the European Academy of Sciences and Arts
1640:
Kumar Chellapilla; Sid Puri; Patrice Simard (2006).
1433:
Simonyan, Karen; Zisserman, Andrew (10 April 2015),
342:over output patterns. The second network learns by
130:
116:
106:
81:
71:
53:
34:
2217:
2215:
1047:IEEE Transactions on Autonomous Mental Development
952:
950:
948:
175:King Abdullah University of Science and Technology
1809:Institute of Electrical and Electronics Engineers
1741:
1739:
1737:
1530:Advances in Neural Information Processing Systems
361:, and more importantly analyzed and overcame the
1947:
1945:
1943:
2319:
2317:
2195:"User Centric AI Creates a New Order for Users"
1389:"Google voice search: faster and more accurate"
1081:Untersuchungen zu dynamischen neuronalen Netzen
821:
819:
704:
702:
2290:
2288:
1616:"Deep Learning: Our Miraculous Year 1990-1991"
1557:. Las Vegas, NV, USA: IEEE. pp. 770–778.
722:
720:
330:that contest with each other in the form of a
1934:Proc. 4th International Conf. Computer Vision
861:
859:
857:
713:. MIT Press/Bradford Books. pp. 222–227.
8:
1550:Deep Residual Learning for Image Recognition
1480:Deep Residual Learning for Image Recognition
1104:Sepp Hochreiter; Jürgen Schmidhuber (1997).
999:
997:
995:
792:
790:
788:
786:
784:
782:
780:
778:
776:
496:, later modified by J. Weng's method called
208:, all of which are widespread in modern AI.
1979:
1977:
387:connectionist temporal classification (CTC)
326:In 1991, Schmidhuber published adversarial
2522:Academic staff of the University of Lugano
2492:German artificial intelligence researchers
295:(RNNs) pre-trained one level at a time by
42:
31:
2137:Recipients: Neural Networks Pioneer Award
2064:Schmidhuber, Juergen (30 December 2022).
2037:Schmidhuber, Juergen (14 December 2023).
1826:
1816:
1763:
1562:
1508:
1487:
1462:
1442:
1357:
1331:
1274:
1221:
1168:
1009:
740:
622:
620:
618:
616:
614:
472:(CNNs) on fast parallel computers called
2168:Heaven, Will Douglas (15 October 2020).
904:
902:
657:on 13 March 2023. Retrieved 9 May 2023.
582:
547:IEEE Computational Intelligence Society
2193:Choul-woong, Yeon (22 February 2023).
2512:Technical University of Munich alumni
1719:Schmidhuber, Jürgen (17 March 2017).
1694:"IJCNN 2011 Competition result table"
1536:. Curran Associates, Inc.: 2377–2385.
551:European Academy of Sciences and Arts
287:with long credit assignment paths in
247:(IDSIA), a Swiss AI lab, since 1995.
165:. He is a scientific director of the
7:
2068:. IDSIA, Switzerland. Archived from
2041:. IDSIA, Switzerland. Archived from
2014:. IDSIA, Switzerland. Archived from
2010:Schmidhuber, Juergen (7 July 2022).
598:
596:
594:
592:
590:
588:
586:
492:'s original CNN architecture called
2091:Fulterer, Ruth (20 February 2021).
1984:Oltermann, Philip (18 April 2017).
157:noted for his work in the field of
153:(born 17 January 1963) is a German
2375:Ruth Fulterer (21 February 2021).
406:for transcription and search, and
234:Università della Svizzera Italiana
25:
2247:Enrique Alpanes (25 April 2021).
886:The Mathematics Genealogy Project
626:John Markoff (27 November 2016).
465:disputes this claim of priority.
400:Google Neural Machine Translation
243:He has served as the director of
224:, Germany. His PhD advisors were
2449:Ivakhnenko, A.G. (March 1970).
1402:Levy, Steven (24 August 2016).
908:Dave O'Leary (3 October 2016).
870:. Springer. pp. 9355–9366.
436:In 1992, Schmidhuber published
381:was published with his student
202:generative adversarial networks
1698:OFFICIAL IJCNN2011 COMPETITION
484:et al. (1989) who applied the
348:generative adversarial network
218:Technical University of Munich
76:Technical University of Munich
1:
2267:Razavi, Hooman (5 May 2020).
1603:. Springer. pp. 460–463.
1526:"Training Very Deep Networks"
797:Jones, Hessie (23 May 2023).
558:Alexey Grigorevich Ivakhnenko
470:convolutional neural networks
2497:Machine learning researchers
2467:10.1016/0005-1098(70)90092-0
2401:Wang, Brian (14 June 2017).
2350:Wong, Andrew (16 May 2018).
1879:; AT&T Bell Laboratories
1811:(IEEE). pp. 3642–3649.
1746:Schmidhuber, Jürgen (2015).
1614:Schmidhuber, Jürgen (2022).
1232:10.1016/j.neunet.2005.06.042
1025:Schmidhuber, Jürgen (1993).
957:Schmidhuber, Jürgen (1992).
751:10.1016/j.neunet.2020.04.008
379:backpropagation through time
2324:Taylor, Josh (7 May 2023).
2295:Colton, Emma (7 May 2023).
1774:10.1162/neco.2006.18.7.1527
319:network into a lower level
190:natural language processing
2543:
2502:German computer scientists
1293:10.1109/TNNLS.2016.2582924
1179:10.1162/089976600300015015
1124:10.1162/neco.1997.9.8.1735
488:algorithm to a variant of
446:feedforward neural network
427:feedforward neural network
363:vanishing gradient problem
289:artificial neural networks
163:artificial neural networks
1837:10.1109/CVPR.2012.6248110
1644:. In Lorette, Guy (ed.).
1059:10.1109/TAMD.2010.2056368
978:10.1162/neco.1992.4.2.234
840:10.1162/neco.1992.4.1.131
442:recurrent neural networks
421:principles to create the
402:, have also been used in
359:neural history compressor
293:recurrent neural networks
126:
99:
41:
27:German computer scientist
2099:(in Swiss High German).
1106:"Long short-term memory"
371:recurrent neural network
340:probability distribution
305:internal representations
297:self-supervised learning
92:, artificial curiosity,
2385:. Accessed August 2021.
2257:. Accessed August 2021.
2222:Dunker, Anders (2020).
1961:Bloomberg Business Week
882:"Jürgen H. Schmidhuber"
438:fast weights programmer
431:residual neural network
194:dynamic neural networks
159:artificial intelligence
111:Artificial intelligence
2142:29 August 2021 at the
2125:INNS Awards Recipients
1891:Biological Cybernetics
1364:Retrieved May 14, 2017
916:. Accessed April 2017.
637:. Accessed April 2017.
524:, who shared the 2018
367:long short-term memory
311:into a single RNN, by
283:did not work well for
186:long short-term memory
86:Long short-term memory
2427:Schmidhuber, Jurgen.
2173:MIT Technology Review
682:"Juergen Schmidhuber"
603:Schmidhuber, Jürgen.
2097:Neue Zürcher Zeitung
1573:10.1109/CVPR.2016.90
1391:. 24 September 2015.
440:, an alternative to
2228:Modern Times Review
2045:on 16 December 2023
1597:Schmidhuber, Jürgen
1285:2015arXiv150304069G
1028:Habilitation Thesis
396:machine translation
18:Juergen Schmidhuber
2018:on 9 February 2023
1903:10.1007/bf00344251
1156:Neural Computation
1111:Neural Computation
1086:2015-03-06 at the
966:Neural Computation
828:Neural Computation
686:scholar.google.com
651:cemse.kaust.edu.sa
647:Jürgen Schmidhuber
605:"Curriculum Vitae"
490:Kunihiko Fukushima
392:speech recognition
369:(LSTM), a type of
365:. This led to the
155:computer scientist
151:Jürgen Schmidhuber
36:Jürgen Schmidhuber
1846:978-1-4673-1226-4
1582:978-1-4673-8851-1
1377:. 11 August 2015.
1269:(10): 2222–2232.
1163:(10): 2451–2471.
1079:S. Hochreiter., "
933:. 16 January 2017
385:in 2005, and its
301:predictive coding
252:self-driving cars
148:
147:
101:Scientific career
16:(Redirected from
2534:
2471:
2470:
2446:
2440:
2439:
2437:
2435:
2424:
2418:
2417:
2415:
2413:
2398:
2392:
2386:
2373:
2367:
2366:
2364:
2362:
2347:
2341:
2340:
2338:
2336:
2321:
2312:
2311:
2309:
2307:
2292:
2283:
2282:
2280:
2278:
2272:Universal Cinema
2264:
2258:
2245:
2239:
2238:
2236:
2234:
2219:
2210:
2209:
2207:
2205:
2190:
2184:
2183:
2181:
2179:
2165:
2159:
2153:
2147:
2134:
2128:
2122:
2116:
2115:
2113:
2111:
2088:
2082:
2081:
2079:
2077:
2061:
2055:
2054:
2052:
2050:
2034:
2028:
2027:
2025:
2023:
2007:
2001:
2000:
1998:
1996:
1981:
1972:
1971:
1969:
1967:
1949:
1938:
1937:
1929:
1923:
1922:
1886:
1880:
1873:
1867:
1866:
1830:
1820:
1807:. New York, NY:
1800:
1794:
1793:
1767:
1743:
1732:
1731:
1729:
1727:
1716:
1710:
1709:
1707:
1705:
1690:
1684:
1683:
1681:
1679:
1665:
1656:
1650:
1649:
1637:
1631:
1630:
1628:
1626:
1611:
1605:
1604:
1593:
1587:
1586:
1566:
1544:
1538:
1537:
1521:
1515:
1514:
1512:
1500:
1494:
1493:
1491:
1475:
1469:
1468:
1466:
1454:
1448:
1447:
1446:
1430:
1424:
1423:
1421:
1419:
1410:. Archived from
1399:
1393:
1392:
1385:
1379:
1378:
1371:
1365:
1363:
1361:
1344:
1338:
1337:
1335:
1319:
1313:
1312:
1278:
1258:
1252:
1251:
1225:
1216:(5–6): 602–610.
1205:
1199:
1198:
1172:
1150:
1144:
1143:
1118:(8): 1735–1780.
1101:
1095:
1077:
1071:
1070:
1042:
1036:
1035:
1033:
1022:
1016:
1015:
1013:
1001:
990:
989:
963:
954:
943:
942:
940:
938:
923:
917:
906:
897:
896:
894:
892:
878:
872:
871:
863:
852:
851:
823:
814:
813:
811:
809:
794:
771:
770:
744:
724:
715:
714:
706:
697:
696:
694:
692:
678:
672:
671:
664:
658:
653:. Archived from
644:
638:
624:
609:
608:
600:
450:gradient descent
444:. It has a slow
344:gradient descent
336:generative model
144:
141:
139:
137:
46:
32:
21:
2542:
2541:
2537:
2536:
2535:
2533:
2532:
2531:
2477:
2476:
2475:
2474:
2448:
2447:
2443:
2433:
2431:
2426:
2425:
2421:
2411:
2409:
2407:Next Big Future
2400:
2399:
2395:
2389:
2374:
2370:
2360:
2358:
2349:
2348:
2344:
2334:
2332:
2323:
2322:
2315:
2305:
2303:
2294:
2293:
2286:
2276:
2274:
2266:
2265:
2261:
2246:
2242:
2232:
2230:
2221:
2220:
2213:
2203:
2201:
2192:
2191:
2187:
2177:
2175:
2167:
2166:
2162:
2154:
2150:
2144:Wayback Machine
2135:
2131:
2123:
2119:
2109:
2107:
2090:
2089:
2085:
2075:
2073:
2072:on 7 April 2023
2063:
2062:
2058:
2048:
2046:
2036:
2035:
2031:
2021:
2019:
2009:
2008:
2004:
1994:
1992:
1983:
1982:
1975:
1965:
1963:
1955:(15 May 2018).
1951:
1950:
1941:
1931:
1930:
1926:
1888:
1887:
1883:
1874:
1870:
1847:
1828:10.1.1.300.3283
1802:
1801:
1797:
1758:(11): 1527–54.
1748:"Deep Learning"
1745:
1744:
1735:
1725:
1723:
1718:
1717:
1713:
1703:
1701:
1692:
1691:
1687:
1677:
1675:
1663:
1658:
1657:
1653:
1639:
1638:
1634:
1624:
1622:
1613:
1612:
1608:
1595:
1594:
1590:
1583:
1546:
1545:
1541:
1523:
1522:
1518:
1502:
1501:
1497:
1477:
1476:
1472:
1456:
1455:
1451:
1432:
1431:
1427:
1417:
1415:
1414:on 23 June 2017
1401:
1400:
1396:
1387:
1386:
1382:
1373:
1372:
1368:
1346:
1345:
1341:
1321:
1320:
1316:
1260:
1259:
1255:
1223:10.1.1.331.5800
1210:Neural Networks
1207:
1206:
1202:
1152:
1151:
1147:
1103:
1102:
1098:
1088:Wayback Machine
1078:
1074:
1044:
1043:
1039:
1031:
1024:
1023:
1019:
1003:
1002:
993:
961:
956:
955:
946:
936:
934:
925:
924:
920:
914:IT World Canada
907:
900:
890:
888:
880:
879:
875:
865:
864:
855:
825:
824:
817:
807:
805:
796:
795:
774:
729:Neural Networks
726:
725:
718:
708:
707:
700:
690:
688:
680:
679:
675:
666:
665:
661:
645:
641:
625:
612:
602:
601:
584:
579:
566:
543:
514:Geoffrey Hinton
512:, in favour of
506:
504:Credit disputes
486:backpropagation
478:computer vision
463:Geoffrey Hinton
448:that learns by
423:highway network
355:Sepp Hochreiter
328:neural networks
315:a higher level
281:backpropagation
277:
256:Sepp Hochreiter
240:, Switzerland.
226:Wilfried Brauer
214:
161:, specifically
134:
72:Alma mater
67:
58:
57:17 January 1963
49:
37:
28:
23:
22:
15:
12:
11:
5:
2540:
2538:
2530:
2529:
2524:
2519:
2514:
2509:
2504:
2499:
2494:
2489:
2479:
2478:
2473:
2472:
2461:(2): 207–219.
2441:
2419:
2393:
2387:
2368:
2342:
2313:
2284:
2259:
2240:
2211:
2199:Korea IT Times
2185:
2160:
2148:
2129:
2117:
2083:
2056:
2029:
2002:
1973:
1939:
1924:
1897:(4): 193–202.
1881:
1868:
1845:
1795:
1765:10.1.1.76.1541
1733:
1711:
1685:
1651:
1632:
1606:
1588:
1581:
1539:
1516:
1495:
1470:
1449:
1425:
1394:
1380:
1366:
1339:
1333:10.1.1.75.6306
1314:
1253:
1200:
1170:10.1.1.55.5709
1145:
1096:
1072:
1053:(3): 230–247.
1037:
1017:
991:
972:(2): 234–242.
944:
918:
898:
873:
853:
834:(1): 131–139.
815:
772:
716:
711:Proc. SAB'1991
698:
673:
659:
639:
634:New York Times
610:
581:
580:
578:
575:
565:
562:
542:
539:
531:New York Times
505:
502:
454:outer products
338:that models a
279:In the 1980s,
276:
273:
269:all-purpose AI
230:Klaus Schulten
213:
210:
146:
145:
132:
128:
127:
124:
123:
118:
114:
113:
108:
104:
103:
97:
96:
83:
82:Known for
79:
78:
73:
69:
68:
59:
55:
51:
50:
47:
39:
38:
35:
26:
24:
14:
13:
10:
9:
6:
4:
3:
2:
2539:
2528:
2525:
2523:
2520:
2518:
2515:
2513:
2510:
2508:
2505:
2503:
2500:
2498:
2495:
2493:
2490:
2488:
2487:Living people
2485:
2484:
2482:
2468:
2464:
2460:
2456:
2452:
2445:
2442:
2430:
2423:
2420:
2408:
2404:
2397:
2394:
2391:
2388:
2384:
2383:
2378:
2372:
2369:
2357:
2353:
2346:
2343:
2331:
2327:
2320:
2318:
2314:
2302:
2298:
2291:
2289:
2285:
2273:
2270:
2263:
2260:
2256:
2255:
2250:
2244:
2241:
2229:
2225:
2218:
2216:
2212:
2200:
2196:
2189:
2186:
2174:
2171:
2164:
2161:
2157:
2152:
2149:
2145:
2141:
2138:
2133:
2130:
2126:
2121:
2118:
2106:
2102:
2098:
2094:
2087:
2084:
2071:
2067:
2060:
2057:
2044:
2040:
2033:
2030:
2017:
2013:
2006:
2003:
1991:
1987:
1980:
1978:
1974:
1962:
1958:
1954:
1953:Vance, Ashlee
1948:
1946:
1944:
1940:
1935:
1928:
1925:
1920:
1916:
1912:
1908:
1904:
1900:
1896:
1892:
1885:
1882:
1878:
1872:
1869:
1864:
1860:
1856:
1852:
1848:
1842:
1838:
1834:
1829:
1824:
1819:
1814:
1810:
1806:
1799:
1796:
1791:
1787:
1783:
1779:
1775:
1771:
1766:
1761:
1757:
1753:
1749:
1742:
1740:
1738:
1734:
1722:
1715:
1712:
1699:
1695:
1689:
1686:
1673:
1669:
1662:
1655:
1652:
1647:
1643:
1636:
1633:
1621:
1617:
1610:
1607:
1602:
1598:
1592:
1589:
1584:
1578:
1574:
1570:
1565:
1560:
1556:
1552:
1551:
1543:
1540:
1535:
1531:
1527:
1520:
1517:
1511:
1506:
1499:
1496:
1490:
1485:
1481:
1474:
1471:
1465:
1460:
1453:
1450:
1445:
1440:
1436:
1429:
1426:
1413:
1409:
1405:
1398:
1395:
1390:
1384:
1381:
1376:
1370:
1367:
1360:
1355:
1351:
1343:
1340:
1334:
1329:
1325:
1318:
1315:
1310:
1306:
1302:
1298:
1294:
1290:
1286:
1282:
1277:
1272:
1268:
1264:
1257:
1254:
1249:
1245:
1241:
1237:
1233:
1229:
1224:
1219:
1215:
1211:
1204:
1201:
1196:
1192:
1188:
1184:
1180:
1176:
1171:
1166:
1162:
1158:
1157:
1149:
1146:
1141:
1137:
1133:
1129:
1125:
1121:
1117:
1113:
1112:
1107:
1100:
1097:
1093:
1089:
1085:
1082:
1076:
1073:
1068:
1064:
1060:
1056:
1052:
1048:
1041:
1038:
1030:
1029:
1021:
1018:
1012:
1007:
1000:
998:
996:
992:
987:
983:
979:
975:
971:
967:
960:
953:
951:
949:
945:
932:
931:Bloomberg.com
928:
922:
919:
915:
911:
905:
903:
899:
887:
883:
877:
874:
869:
862:
860:
858:
854:
849:
845:
841:
837:
833:
829:
822:
820:
816:
804:
800:
793:
791:
789:
787:
785:
783:
781:
779:
777:
773:
768:
764:
760:
756:
752:
748:
743:
738:
734:
730:
723:
721:
717:
712:
705:
703:
699:
687:
683:
677:
674:
669:
663:
660:
656:
652:
648:
643:
640:
636:
635:
629:
623:
621:
619:
617:
615:
611:
606:
599:
597:
595:
593:
591:
589:
587:
583:
576:
574:
570:
563:
561:
559:
554:
552:
548:
540:
538:
534:
532:
527:
523:
519:
518:Yoshua Bengio
515:
511:
510:deep learning
503:
501:
499:
495:
491:
487:
483:
479:
475:
471:
466:
464:
460:
455:
451:
447:
443:
439:
434:
432:
428:
424:
420:
416:
411:
409:
405:
401:
397:
393:
388:
384:
380:
376:
372:
368:
364:
360:
356:
351:
349:
345:
341:
337:
333:
332:zero-sum game
329:
324:
322:
318:
314:
310:
306:
302:
298:
294:
290:
286:
285:deep learning
282:
274:
272:
270:
265:
264:Marcus Hutter
261:
257:
253:
248:
246:
241:
239:
235:
231:
227:
223:
219:
211:
209:
207:
203:
199:
198:meta-learning
195:
191:
187:
182:
180:
176:
172:
168:
164:
160:
156:
152:
143:
133:
129:
125:
122:
119:
115:
112:
109:
105:
102:
98:
95:
94:meta-learning
91:
90:Gödel machine
87:
84:
80:
77:
74:
70:
66:
62:
56:
52:
45:
40:
33:
30:
19:
2458:
2454:
2444:
2432:. Retrieved
2422:
2410:. Retrieved
2406:
2396:
2390:
2380:
2371:
2359:. Retrieved
2355:
2345:
2333:. Retrieved
2330:The Guardian
2329:
2304:. Retrieved
2300:
2275:. Retrieved
2271:
2262:
2252:
2243:
2231:. Retrieved
2227:
2202:. Retrieved
2198:
2188:
2176:. Retrieved
2172:
2163:
2151:
2132:
2120:
2108:. Retrieved
2096:
2086:
2074:. Retrieved
2070:the original
2059:
2047:. Retrieved
2043:the original
2032:
2020:. Retrieved
2016:the original
2005:
1993:. Retrieved
1990:The Guardian
1989:
1964:. Retrieved
1960:
1933:
1927:
1894:
1890:
1884:
1871:
1804:
1798:
1755:
1752:Scholarpedia
1751:
1724:. Retrieved
1714:
1702:. Retrieved
1697:
1688:
1676:. Retrieved
1671:
1667:
1654:
1645:
1635:
1623:. Retrieved
1619:
1609:
1600:
1596:
1591:
1554:
1549:
1542:
1533:
1529:
1519:
1498:
1479:
1473:
1452:
1434:
1428:
1416:. Retrieved
1412:the original
1407:
1397:
1383:
1369:
1342:
1323:
1317:
1266:
1262:
1256:
1213:
1209:
1203:
1160:
1154:
1148:
1115:
1109:
1099:
1091:
1075:
1050:
1046:
1040:
1027:
1020:
969:
965:
935:. Retrieved
930:
921:
913:
889:. Retrieved
885:
876:
867:
831:
827:
806:. Retrieved
802:
732:
728:
710:
689:. Retrieved
685:
676:
668:"Leadership"
662:
655:the original
650:
642:
631:
571:
567:
555:
544:
535:
530:
526:Turing Award
507:
494:neocognitron
467:
459:Transformers
437:
435:
412:
404:Google Voice
358:
352:
325:
320:
316:
308:
278:
260:Jaan Tallinn
249:
242:
215:
206:transformers
183:
179:Saudi Arabia
150:
149:
117:Institutions
100:
65:West Germany
29:
2527:1963 births
2434:26 December
2412:27 February
2361:27 February
2110:19 December
2049:19 December
1995:23 February
1678:17 November
1674:: 1237–1242
1648:. Suvisoft.
1326:: 369–376.
937:23 February
541:Recognition
498:max-pooling
383:Alex Graves
321:automatizer
204:and linear
177:(KAUST) in
171:Switzerland
2481:Categories
2455:Automatica
1966:16 January
1936:: 121–128.
1726:14 January
1704:14 January
1601:ICANN 1993
1564:1512.03385
1510:1505.00387
1489:1512.03385
1464:1502.01852
1359:1609.08144
1350:Dean, Jeff
1276:1503.04069
1011:2212.11279
742:1906.04493
691:20 October
577:References
522:Yann LeCun
482:Yann LeCun
375:Felix Gers
313:distilling
299:. It uses
2277:20 August
2233:20 August
2178:20 August
2105:0376-6829
1919:206775608
1855:812295155
1823:CiteSeerX
1818:1202.2745
1760:CiteSeerX
1444:1409.1556
1328:CiteSeerX
1218:CiteSeerX
1165:CiteSeerX
868:ICML 2021
767:216056336
735:: 58–66.
309:collapsed
303:to learn
142:/~juergen
2301:Fox News
2140:Archived
1782:16764513
1620:idsia.ch
1301:27411231
1240:16112549
1195:11598600
1187:11032042
1084:Archived
986:18271205
848:16683347
759:32334341
415:training
275:Research
2254:El Pais
2156:Members
1911:7370364
1863:2161592
1790:2309950
1625:23 July
1418:23 June
1309:3356463
1281:Bibcode
1248:1856462
1140:1915014
1132:9377276
1094:, 1991.
317:chunker
131:Website
2335:26 May
2306:26 May
2204:26 May
2103:
1917:
1909:
1861:
1853:
1843:
1825:
1788:
1780:
1762:
1700:. 2010
1579:
1330:
1307:
1299:
1246:
1238:
1220:
1193:
1185:
1167:
1138:
1130:
1067:234198
1065:
984:
891:5 July
846:
808:26 May
803:Forbes
765:
757:
461:, but
262:, and
238:Lugano
222:Munich
212:Career
138:.idsia
136:people
107:Fields
61:Munich
2076:3 May
2022:3 May
1915:S2CID
1859:S2CID
1813:arXiv
1786:S2CID
1664:(PDF)
1559:arXiv
1505:arXiv
1484:arXiv
1459:arXiv
1439:arXiv
1408:Wired
1354:arXiv
1305:S2CID
1271:arXiv
1244:S2CID
1191:S2CID
1136:S2CID
1063:S2CID
1032:(PDF)
1006:arXiv
982:S2CID
962:(PDF)
844:S2CID
763:S2CID
737:arXiv
564:Views
2436:2019
2414:2019
2363:2019
2356:CNBC
2337:2023
2308:2023
2279:2021
2235:2021
2206:2023
2180:2021
2112:2023
2101:ISSN
2078:2023
2051:2023
2024:2023
1997:2018
1968:2019
1907:PMID
1851:OCLC
1841:ISBN
1778:PMID
1728:2019
1706:2019
1680:2013
1627:2024
1577:ISBN
1420:2017
1297:PMID
1236:PMID
1183:PMID
1128:PMID
939:2018
893:2022
810:2023
755:PMID
693:2021
632:The
520:and
474:GPUs
425:, a
419:LSTM
408:Siri
394:and
228:and
54:Born
2463:doi
2382:NZZ
1899:doi
1833:doi
1770:doi
1569:doi
1289:doi
1228:doi
1175:doi
1120:doi
1090:,"
1055:doi
974:doi
836:doi
747:doi
733:127
236:in
220:in
169:in
140:.ch
2483::
2457:.
2453:.
2405:.
2379:.
2354:.
2328:.
2316:^
2299:.
2287:^
2251:.
2226:.
2214:^
2197:.
2095:.
1988:.
1976:^
1959:.
1942:^
1913:.
1905:.
1895:36
1893:.
1857:.
1849:.
1839:.
1831:.
1821:.
1784:.
1776:.
1768:.
1756:10
1754:.
1750:.
1736:^
1696:.
1670:.
1666:.
1618:.
1575:.
1567:.
1553:.
1534:28
1532:.
1528:.
1482:.
1437:,
1406:.
1303:.
1295:.
1287:.
1279:.
1267:28
1265:.
1242:.
1234:.
1226:.
1214:18
1212:.
1189:.
1181:.
1173:.
1161:12
1159:.
1134:.
1126:.
1114:.
1108:.
1061:.
1049:.
994:^
980:.
968:.
964:.
947:^
929:.
912:.
901:^
884:.
856:^
842:.
830:.
818:^
801:.
775:^
761:.
753:.
745:.
731:.
719:^
701:^
684:.
649:.
630:.
613:^
585:^
553:.
516:,
500:.
410:.
258:,
254:.
200:,
196:,
181:.
88:,
63:,
2469:.
2465::
2459:6
2438:.
2416:.
2365:.
2339:.
2310:.
2281:.
2237:.
2208:.
2182:.
2114:.
2080:.
2053:.
2026:.
1999:.
1970:.
1921:.
1901::
1865:.
1835::
1815::
1792:.
1772::
1730:.
1708:.
1682:.
1672:2
1629:.
1585:.
1571::
1561::
1513:.
1507::
1492:.
1486::
1467:.
1461::
1441::
1422:.
1362:.
1356::
1336:.
1311:.
1291::
1283::
1273::
1250:.
1230::
1197:.
1177::
1142:.
1122::
1116:9
1069:.
1057::
1051:2
1034:.
1014:.
1008::
988:.
976::
970:4
941:.
895:.
850:.
838::
832:4
812:.
769:.
749::
739::
695:.
670:.
607:.
20:)
Text is available under the Creative Commons Attribution-ShareAlike License. Additional terms may apply.