作者: reject 時(shí)間: 2025-3-21 23:24 作者: 摘要記錄 時(shí)間: 2025-3-22 02:17
A minimax lower bound for empirical quantizer design, empirically designed vector quantizer is at least . (..) away from the optimal distortion for some distribution on a bounded subset of .., where . is the number of i.i.d. data points that are used to train the empirical quantizer.作者: 可用 時(shí)間: 2025-3-22 05:15 作者: Bone-Scan 時(shí)間: 2025-3-22 11:58
On learning branching programs and small depth circuits,We study the learnability of branching programs and small-depth circuits with modular and threshold gates in both the exact and PAC learning models with and without membership queries. Our results extend earlier works [11, 18, 15] and exhibit further applications of . [7] in learning theory.作者: 無彈性 時(shí)間: 2025-3-22 14:29
Computational Learning Theory978-3-540-68431-2Series ISSN 0302-9743 Series E-ISSN 1611-3349 作者: 無彈性 時(shí)間: 2025-3-22 19:35 作者: garrulous 時(shí)間: 2025-3-22 23:06
Spezielle Varianten der SEM-Analyse, empirically designed vector quantizer is at least . (..) away from the optimal distortion for some distribution on a bounded subset of .., where . is the number of i.i.d. data points that are used to train the empirical quantizer.作者: sed-rate 時(shí)間: 2025-3-23 05:24 作者: verdict 時(shí)間: 2025-3-23 05:51 作者: Veneer 時(shí)間: 2025-3-23 12:36 作者: 窗簾等 時(shí)間: 2025-3-23 16:01
Zielsetzung und Problemstellung,et al. [7], it is assumed that membership queries on instances near the boundary of the target concept may receive a “don‘t know” answer..We show that zero-one threshold functions are efficiently learnable in this model. The learning algorithm uses split graphs when the boundary region has radius 1,作者: 聲明 時(shí)間: 2025-3-23 20:13
Zusammenfassung und Diskussion,m of machine learning using randomly drawn examples. Quite often in practice some form of . partial information about the target is available in addition to randomly drawn examples. In this paper we extend the PAC model to a scenario of learning with partial information in addition to randomly drawn作者: ACE-inhibitor 時(shí)間: 2025-3-23 23:49 作者: Figate 時(shí)間: 2025-3-24 03:54 作者: 較早 時(shí)間: 2025-3-24 06:48 作者: 令人悲傷 時(shí)間: 2025-3-24 12:33 作者: endocardium 時(shí)間: 2025-3-24 14:50
Diskussion ausgew?hlter Beispieleandomized hypotheses for learning with small sample sizes and high malicious noise rates. We show an algorithm that PAC learns any target class of VC-dimension . using randomized hypotheses and order of . training examples (up to logarithmic factors) while tolerating malicious noise rates even sligh作者: 膽大 時(shí)間: 2025-3-24 19:13 作者: MAIZE 時(shí)間: 2025-3-25 02:12 作者: 鞭打 時(shí)間: 2025-3-25 03:33
https://doi.org/10.1007/978-3-658-09609-0 asking for the value of . at adaptively chosen inputs. For classes . consisting of functions which may be obtained from one function . on . inputs by replacing arbitrary . inputs by given constants this problem is known as attribute-efficient learning with . essential attributes. Results on general作者: RADE 時(shí)間: 2025-3-25 08:11
https://doi.org/10.1007/978-3-658-09609-0 most . times. In particular, we abbreviate . for .=1. The language . of a pattern . is the set of all strings obtained by substituting any non-null constant string for each variable symbol in .. In this paper, we show that any pattern . is exactly identifiable in .(|.|.) time from one positive exam作者: 眉毛 時(shí)間: 2025-3-25 15:36
,Datenqualit?t und Messmodelle,..,A. in . such that the sum of the elements contained in the subarrays is maximum..In this paper, we present a simple algorithm, based on Dynamic Programming, solving SUB(.) in time .. Extracting the main idea of the dynamic programming scheme, we are able to extend the algorithm such that it is ap作者: 難聽的聲音 時(shí)間: 2025-3-25 17:59
Spezielle Varianten der SEM-Analyse, empirically designed vector quantizer is at least . (..) away from the optimal distortion for some distribution on a bounded subset of .., where . is the number of i.i.d. data points that are used to train the empirical quantizer.作者: 托運(yùn) 時(shí)間: 2025-3-25 20:15 作者: certitude 時(shí)間: 2025-3-26 01:08
https://doi.org/10.1007/3-540-62685-9Algorithmische Komplexit?t; Algorithmisches Lernen; Induktive Inferenz; Neuronale Netze; Sprachenlernen; 作者: PRO 時(shí)間: 2025-3-26 05:14 作者: 半球 時(shí)間: 2025-3-26 11:07
https://doi.org/10.1007/978-3-658-09609-0 reduced to |.|+|.| if the target pattern . is not critical. For instance, any pattern . whose constant parts are of length at most 3 is not critical. Finally, we show a nontrivial subclass of . that is identified using membership queries only, without any initial positive example.作者: 發(fā)現(xiàn) 時(shí)間: 2025-3-26 14:58
learning pattern languages using queries, reduced to |.|+|.| if the target pattern . is not critical. For instance, any pattern . whose constant parts are of length at most 3 is not critical. Finally, we show a nontrivial subclass of . that is identified using membership queries only, without any initial positive example.作者: chisel 時(shí)間: 2025-3-26 20:47 作者: 不斷的變動(dòng) 時(shí)間: 2025-3-26 21:16 作者: Kaleidoscope 時(shí)間: 2025-3-27 01:38
Zielsetzung und Problemstellung, and their generalization to split hypergraphs (for which we give a split-finding algorithm) when the boundary region has constant radius greater than 1. We use a notion of indistinguishability of concepts that is appropriate for this model.作者: CLEFT 時(shí)間: 2025-3-27 06:20
Progress in Colloid and Polymer Sciencenerate boxes (that is, boxes that depend only on .(1) variables) in time poly(., log ?) (generalizing the learnability of .(1)-DNF and of boxes in .(1) dimensions). The algorithm for this class uses only equivalence queries and it can also be used to learn the class of unions of .(1) boxes (from equivalence queries only).作者: 擁護(hù)者 時(shí)間: 2025-3-27 12:24 作者: 者變 時(shí)間: 2025-3-27 16:07 作者: 大約冬季 時(shí)間: 2025-3-27 21:19
Learning boxes in high dimension,nerate boxes (that is, boxes that depend only on .(1) variables) in time poly(., log ?) (generalizing the learnability of .(1)-DNF and of boxes in .(1) dimensions). The algorithm for this class uses only equivalence queries and it can also be used to learn the class of unions of .(1) boxes (from equivalence queries only).作者: breadth 時(shí)間: 2025-3-27 22:56 作者: 寡頭政治 時(shí)間: 2025-3-28 02:16 作者: Herd-Immunity 時(shí)間: 2025-3-28 10:20 作者: 不可救藥 時(shí)間: 2025-3-28 11:05
Learning from incomplete boundary queries using split graphs and hypergraphs, and their generalization to split hypergraphs (for which we give a split-finding algorithm) when the boundary region has constant radius greater than 1. We use a notion of indistinguishability of concepts that is appropriate for this model.作者: LIMN 時(shí)間: 2025-3-28 16:15 作者: 增減字母法 時(shí)間: 2025-3-28 22:28 作者: 打谷工具 時(shí)間: 2025-3-28 23:48
Die Gemeinde in der Regionalen Verflechtunghanging continuously at a constant rate from one extreme distribution to another. We show in this case how to use a simple weighting scheme to estimate the error of an hypothesis, and using this estimate, to minimize the error of the prediction.作者: 平庸的人或物 時(shí)間: 2025-3-29 03:22 作者: Rejuvenate 時(shí)間: 2025-3-29 09:50
https://doi.org/10.1007/978-3-658-09609-0 replacing arbitrary . inputs by given constants this problem is known as attribute-efficient learning with . essential attributes. Results on general classes of functions are known. More precise and often optimal results are presented for the cases where . is one of the functions disjunction, parity or threshold.作者: 邊緣 時(shí)間: 2025-3-29 13:32
Learning under persistent drift,hanging continuously at a constant rate from one extreme distribution to another. We show in this case how to use a simple weighting scheme to estimate the error of an hypothesis, and using this estimate, to minimize the error of the prediction.作者: 大門在匯總 時(shí)間: 2025-3-29 18:00 作者: CLOWN 時(shí)間: 2025-3-29 22:57 作者: 紅潤(rùn) 時(shí)間: 2025-3-30 03:41
Conference proceedings 1997el, in March 1997..The book presents 25 revised full papers carefully selected from a total of 36 high-quality submissions. The volume spans the whole spectrum of computational learning theory, with a certain emphasis on mathematical models of machine learning. Among the topics addressed are machine作者: Meditative 時(shí)間: 2025-3-30 04:45
Learning boxes in high dimension,of .(log .) boxes in time poly(., log ?) (solving an open problem of [15, 11]). (2) The class of unions of disjoint boxes in time poly(.,log ?), where . is the number of boxes. (Previously this was known only in the case where all boxes are disjoint in one of the dimensions). In particular our algor作者: Cumulus 時(shí)間: 2025-3-30 10:29 作者: linguistics 時(shí)間: 2025-3-30 13:48 作者: 王得到 時(shí)間: 2025-3-30 20:36
Learning from incomplete boundary queries using split graphs and hypergraphs,et al. [7], it is assumed that membership queries on instances near the boundary of the target concept may receive a “don‘t know” answer..We show that zero-one threshold functions are efficiently learnable in this model. The learning algorithm uses split graphs when the boundary region has radius 1,作者: gerrymander 時(shí)間: 2025-3-30 22:35 作者: GRAVE 時(shí)間: 2025-3-31 02:10
Monotonic and dual-monotonic probabilistic language learning of indexed families with high probabilive data. In particular, we consider the special case where the probability is equal to 1..Earlier results in the field of probabilistic identification established that — considering function identification — each collection of recursive functions identifiable with probability .>1/2 is deterministic作者: 使?jié)M足 時(shí)間: 2025-3-31 08:35 作者: 痛苦一下 時(shí)間: 2025-3-31 12:54 作者: Allure 時(shí)間: 2025-3-31 13:35
Learning under persistent drift,re the changes might be rapid but their “direction” is relatively constant. We model this type of change by assuming that the target distribution is changing continuously at a constant rate from one extreme distribution to another. We show in this case how to use a simple weighting scheme to estimat作者: arthroplasty 時(shí)間: 2025-3-31 18:36
Randomized hypotheses and minimum disagreement hypotheses for learning with noise,andomized hypotheses for learning with small sample sizes and high malicious noise rates. We show an algorithm that PAC learns any target class of VC-dimension . using randomized hypotheses and order of . training examples (up to logarithmic factors) while tolerating malicious noise rates even sligh作者: 拖網(wǎng) 時(shí)間: 2025-3-31 22:32
Learning when to trust which experts,hat this assumption does not take advantage of situations where both the outcome and the experts‘ predictions are based on some input which the learner gets to observe too. In particular, we exhibit a situation where each individual expert performs badly but collectively they perform well, and show 作者: Dorsal 時(shí)間: 2025-4-1 05:13 作者: 火光在搖曳 時(shí)間: 2025-4-1 06:04 作者: 音的強(qiáng)弱 時(shí)間: 2025-4-1 11:01 作者: Basilar-Artery 時(shí)間: 2025-4-1 16:14 作者: PSA-velocity 時(shí)間: 2025-4-1 19:38
A minimax lower bound for empirical quantizer design, empirically designed vector quantizer is at least . (..) away from the optimal distortion for some distribution on a bounded subset of .., where . is the number of i.i.d. data points that are used to train the empirical quantizer.作者: GULP 時(shí)間: 2025-4-2 01:55
Vapnik-Chervonenkis dimension of recurrent neural networks,o widely used in learning applications, in particular when time is a relevant parameter. This paper provides lower and upper bounds for the VC dimension of such networks. Several types of activation functions are discussed, including threshold, polynomial, piecewise-polynomial and sigmoidal function
