D. Angluin. On the complexity of minimum inference of regular sets. Information and Control, 39(3):337--350, 1978.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....associated with the set of experiences in Figure 3b. Notice that the causal and view graphs associated with the experiences in Figure 3a are isomorphic. The problem of distinguishing environment states by outputs (views) and inputs (actions) has been studied in the framework of automata theory [Angluin, 1978, Gold, 1978, Rivest and Schapire, 1987, Basye et al. 1995] In this framework, the problem we address is the one of finding the minimum automaton (w.r.t. the number of states) consistent with a given set of input output pairs. Without any particular assumptions about the environment or the ....

....we address is the one of finding the minimum automaton (w.r.t. the number of states) consistent with a given set of input output pairs. Without any particular assumptions about the environment or the agent s perceptual abilities, the problem of finding this smallest automaton is NP complete ([Angluin, 1978, Gold, 1978] 5 Topological maps Actions in the causal theory convey patterns of experience but not spatial configuration. Spatial configuration is considered by the topological theory where actions are categorized into two classes: turns and travels. Turns and travels are explained by a new ....

Dana Angluin. On the complexity of minimum inference of regular sets. Information and Control, 39:337--350, 1978.

....Theorem 3 Let E be a complete set of experiences. Then, ceq(ds, ds2) A, v [View(A(ds) v) View(A(ds2) v) Proof. See appendix E (page 234) The problem of distinguishing environment states by outputs (views) and inputs (actions) has been studied in the framework of automata theory [Angluin, 1978, Gold, 1978, Rivest and Schapire, 1987, Basye et al. 1995] In this framework, the problem we address is the one of finding the minimum automaton (w.r.t. the number of states) consistent with a 28 given set of input output pairs. Without any particular assumptions about the environment or the ....

....we address is the one of finding the minimum automaton (w.r.t. the number of states) consistent with a 28 given set of input output pairs. Without any particular assumptions about the environment or the agent s perceptual abilities, the problem of finding this smallest automaton is NP complete ([Angluin, 1978, Gold, 1978] The reader is referred to [Basye et al. 1995] for an example of how to use automata to model dynamical systems. Consider the set of experiences E as in example 2 (page 23) see figure 4.3a) Action type(ml, travel) CS(M, a, ral, b) CS(s2, b, ral, c) Uiew(a, v ) Uiew, ....

[Article contains additional citation context not shown here]

Dana Angluin. On the complexity of minimum inference of regular sets. Information and Control, 39:337-350, 1978.

....The robot gets information about the automaton by actively experimenting in the environment and by observing input output behavior. Rivest and Schapire show that a robot with a teacher can with high probability learn such an environment. They use homing sequences to improve Angluin s algorithm [1] to learn without using a reset mechanism. Ron and Rubinfeld [71] further extend this result by giving an efficient algorithm that with high probability learns finite automata with small cover time, without requiring a teacher. Dean et al. 33] study the problem of learning finite automaton when ....

Dana Angluin. On the complexity of minimum inference of regular sets. Information and Computation, 39:337--350, 1987.

....3 Let E be a complete set of experiences. Then, ceq#ds # ;ds # # ##A; v #V iew#A#ds # #;v# # V iew#A#ds # #;v## : Proof. See appendix E (page 234) The problem of distinguishing environment states by outputs (views) and inputs (actions) has been studied in the framework of automata theory [Angluin, 1978, Gold, 1978, Rivest and Schapire, 1987, Basye et al. 1995] In this framework, the problem we address is the one of finding the minimum automaton (w.r.t. the number of states) consistent with a given set of input output pairs. Without any particular assumptions about the environment or the ....

....we address is the one of finding the minimum automaton (w.r.t. the number of states) consistent with a given set of input output pairs. Without any particular assumptions about the environment or the agent s perceptual abilities, the problem of finding this smallest automaton is NPcomplete ([Angluin, 1978, Gold, 1978] The reader is referred to [Basye et al. 1995] for an example of how to use automata to model dynamical systems. Consider the set of experiences E as in example 2 (page 23) see figure 4.3a) CS#s#;a;ml;b# ; CS#s#;b;ml;c# ; Since the same view is experienced at a, b and c, ....

[Article contains additional citation context not shown here]

Dana Angluin. On the complexity of minimum inference of regular sets. Information and Control, 39:337--350, 1978.

....## #ds; a; ds # # #. where DS#E#=ceq denotes the set of equivalence classes of DS(E) modulo ceq,and #ds# denotes the equivalence class of ds given ceq. The problem of distinguishing environment states by outputs (views) and inputs (actions) has been studied in the framework of automata theory [Angluin, 1978, Gold, 1978, Rivest and Schapire, 1987, Basye et al. 1995] In this framework, the problem we address is the one of finding the minimum automaton (w.r.t. the number of states) consistent with a given set of input output pairs. Without any particular assumptions about the environment or the ....

....we address is the one of finding the minimum automaton (w.r.t. the number of states) consistent with a given set of input output pairs. Without any particular assumptions about the environment or the agent s perceptual abilities, the problem of finding this smallest automaton is NP complete ([Angluin, 1978, Gold, 1978] DRAFT November 13, 2001 12 a,b,c ml a b c d ml ml ml ml ml v v a b c Figure 5: a) b) Causal graphs associated with the set of experiences in Figures 3a and 3b. c) view graph associated with the set of experiences in figure 3b. Notice that the causal ....

Dana Angluin. On the complexity of minimum inference of regular sets. Information and Control, 39:337--350, 1978.

.... from the countable subset of the ultimately periodic words (of type uv w , u and v being finite words) Saoudi and Yokomori [12] define a (restricted) class of local languages, and prove the learnability of these languages from positive examples; Maler and Pnueli [9] adapt Angluin s L algorithm [2] and make it possible to learn a particular class of automata with the assistance of a polynomial number of equivalence and membership queries. Nevertheless, we wish the learning of an automaton to be done from experimental data received from the potential users of a system. The data will ....

D. Angluin. On the Complexity of Minimum Inference of Regular Sets. Information and Control 39, 337--350, 1978.

....and in the length of the longest counterexample supplied by the teacher. Note that Angluin s procedure depends critically on the availability of a reset. The problem of learning an automaton by passively observing its behavior is now well established to be a hard computational problem. Angluin [1] and Gold [16] proved that finding an automaton of n states or less agreeing with a given sample of input output pairs is NP complete. Pitt and Warmuth [22] showed that merely finding an approximate solution is infeasible if P 6= NP. Kearns and Valiant [18] extending the work of Pitt and Warmuth ....

Dana Angluin. On the complexity of minimum inference of regular sets. Information and Control, 39:337--350, 1978. 34

....and the target functions give di#erent results. For example, Angluin [2] has shown that a polynomial number of membership and equivalence queries are su#cient to exactly identify any regular set (finite automaton) whereas the problem of learning a regular set from random examples is NP complete [1]. Even if information is gathered from random examples, cryptanalytic learning scenarios may also vary in the prior knowledge available to the attacker learner about the distribution of those examples. For example, some cryptosystems can be successfully attacked with only general knowledge of the ....

....in 2 CNF (CNF with two literals per clause, which properly contains the set of functions representable in 2 term DNF) is tractable, and so PAC learning the class of functions representable in 2 term DNF is possible, as long as the learner may output his answers in 2 CNF. Similarly, Angluin [1] has shown that it is NP complete to find a minimum size DFA that is consistent with a given set of input output examples, and Pitt and Warmuth [33] have extended this result to show that finding an approximately minimum size DFA consistent with such a set of examples is impossible to do ....

Dana Angluin. On the complexity of minimum inference of regular sets. Information and Control, 39:337--350, 1978.

....in the language which represent most of the probability mass of the language, which makes the task somewhat easier. We also want the learner to identify the probabilities of those strings, which makes the task considerably harder. There s no such thing as a free lunch. On the more practical side, [10] and [11] showed that learning the smallest DFA consistent with a set of training data is NP complete. Even learning a DFA whose size is a polynomial of the smallest DFA is NP hard, under the assumption P 6= NP, 12] and not PAC learnable, under various cryptographic assumptions [13] Since a ....

Dana Angluin, "On the Complexity of Minimum Inference of Regular Sets," Information and Control 39 (1978), 337--350.

No context found.

Angluin, Dana 1978. On the complexity of minimum inference of regular sets. Information and Control 39:337--350. 33

No context found.

D. Angluin. On the complexity of minimum inference of regular sets. Information and Control, 39(3):337--350, 1978.

No context found.

D. Angluin. On the complexity of minimum inference of regular sets. Information and Control, 39:337--350, 1978.

No context found.

Dana Angluin. On the complexity of minimum inference of regular sets. Information and Control, 39:337--350, 1978.

No context found.

D. Angluin, On the complexity of minimum inference of regular sets, Inform. and Control 39 (1978) 337-- 350.

No context found.

Angluin, D. (1978). On the complexity of minimum inference of regular sets. Information and Control, 39:337--350.

No context found.

D. Angluin. On the complexity of minimum inference of regular sets. Information and Control, 39:337--350, 1978.

No context found.

D. Angluin. On the complexity of minimum inference of regular sets. Information and Control, pages 337--350, 1978.

No context found.

D. Angluin. On the complexity of minimum inference of regular sets. Inform. Control, 39(3):337--350, 1978.

No context found.

D. Angluin. On the complexity of minimum inference of regular sets. Information and Control, 39(3):337-350, 1978.

No context found.

Dana Angluin. On the complexity of minimum inference of regular sets. Information and Control, 39:337-350, 1978.

No context found.

Angluin, D. (1978). On the complexity of minimum inference of regular sets. Information and Control, 29(3), 741-765.

No context found.

D. Angluin. On the complexity of minimum inference of regular sets. Information and Control, 39:337-350, 1978.

No context found.

Dana Angluin. On the complexity of minimum inference of regular sets. Information and Control, 39:337--350, 1978.

No context found.

D. Angluin. "On the complexity of minimum inference of regular sets", Information and Control 39 (1978), 302-320.

No context found.

D. Angluin. "On the complexity of minimum inference of regular sets", Information and Control 39 (1978), 302--320. 16

First 50 documents  Next 50

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC