Playing chess
50
0
0
Texte intégral
(2) Playing chess Who is the cleverest?. 2.
(3) Playing with chess Who is the cleverest?. Baron Wolfgang von Kempelen (1769) 3.
(4) Amoeba How to use an incomplete and uncertain model of the environment to perceive, infer, decide and act efficiently enough to survive ?. 4.
(5) Overview How to survive (perceive, reason, learn, decide and act) with incomplete information ? Probability as an alternative to logic How to develop better artifacts using Bayesian reasoning? Biological plausibility of Bayesian reasoning at a macroscopic level? Biological plausibility of Bayesian reasoning at a microscopic level?. 5.
(6) Overview How to survive (perceive, reason, learn, decide and act) with incomplete information ? Probability as an alternative to logic How to develop better artifacts using Bayesian reasoning? Biological plausibility of Bayesian reasoning at a macroscopic level? Biological plausibility of Bayesian reasoning at a microscopic level?. 6.
(7) Probability as alternative to logic Incompleteness Preliminary Knowledge + Experimental Data = Probabilistic Representation. Learning Entropy Principles. Uncertainty P(a) + P(¬a) = 1. Bayesian inference. P(a ∧b) = P(a) × P(b | a) = P(b) × P(a | b). Decision 7.
(8) Bayesian Programming & ProBT® main () {. //Variables Specification. plFloat read_time; plIntegerType id_type(0,1); plFloat times[5] = {1,2,3,5,10}; plSparseType time_type(5,times); plSymbol id("id",id_type); plSymbol time("time",time_type);. //Parametrical forms. Description. Bayesian Program. •Variables. •Decomposition. //Construction of P(id) plProbValue id_dist[2] = {0.75,0.25}; plProbTable P_id(id,id_dist); //Construction of P(time | id = john) plProbValue t_john_dist[5] = {20,30,10,5,2}; plProbTable P_t_john(time,t_john_dist); //Construction of P(time | id = bill) plProbValue t_bill_dist[5] = {2,6,10,40,20}; plProbTable P_t_bill(time,t_bill_dist);. •Parametric Forms. //Construction de P(time | id) plKernelTable Pt_id(time,id); plValues t_and_id(time^id); t_and_id[id] = 0; Pt_id.push(P_t_john,t_and_id); t_and_id[id] = 1;. Identification. Pt_id.push(P_t_bill,t_and_id);. //Decomposition from instances •Learning. // P(time id) = P(id) P(time | id) plJointDistribution jd(time^id,P_id*Pt_id);. Question. //Question. //Getting the question P(id | time) plCndKernel Pid_t; jd.ask(Pid_t,id,time);. ProBAYES.com Bayesian-Programming.org //Read a time from the key board cout<<"P(id,time)= "<<Pid_t<<"\n"; cout<<"Time? : "; cin>>read_time; //Getting P(id | time = read_time) plKernel Pid_readTime;.
(9) Overview How to survive (perceive, reason, learn, decide and act) with incomplete information ? Probability as an alternative to logic How to develop better artifacts using Bayesian reasoning? Biological plausibility of Bayesian reasoning at a macroscopic level? Biological plausibility of Bayesian reasoning at a microscopic level?. 9.
(10) Robotics. PhD Olivier Lebeltel. PhD Kamel Mekhnacha. PhD Ruben Garcia. PhD Christophe Coué. PhD Cédric Pradalier. PhD Carla Koike. PhD Ronan Le Hy. 10. ProBAYES.
(11) Overview How to survive (perceive, reason, learn, decide and act) with incomplete information ? Probability as an alternative to logic How to develop better artifacts using Bayesian reasoning? Biological plausibility of Bayesian reasoning at a macroscopic level? Biological plausibility of Bayesian reasoning at a microscopic level?. 11.
(12) Modeling behaviors. 288. PhD Jihene Serkhane. Laurens Jean and Droulez Jacques PhD C. Moulin-Frier. vier Lebeltel. vier Lebeltel. PhD Francis Colas. PhD Jean Laurens. 12. PostDoc Francis Colas.
(13) Bayesian Action Perception: Handwriting experiments. PhD Estelle Gilet. 13.
(14) Motor Equivalence?. 14.
(15) Motor Equivalence?. • Writer “style” [Wright90]. • Common activated motor areas [Wing00]. [Serratrice93]. 15.
(16) Simulation of action during perception?. [Calvo-Merino04] 16.
(17) Simulation of action during perception? [Longcamp03]. Writing. Pseudo letter reading. Letter reading 17.
(18) Reading. • OCR [Meulenbroek96] [Flash95]. • Human models [Crettez98] [Vuori02] [Dehaene07]. 18.
(19) Writing. • [Hinton05] • [Meulenbroek96] • [Flash95]. 19.
(20) BAP model. 20.
(21) A common space for motor and perception internal representation. 21.
(22) Common features for both representations. 22.
(23) Here come the probabilities. 23.
(24) Learning succession of control points. 24.
(25) Learning succession of control points. 25.
(26) Learning succession of control points. 26.
(27) Learning succession of control points. 27.
(28) Learning succession of control points. 28.
(29) BAP model. 29.
(30) BAP model. 30.
(31) BAP model. 31.
(32) BAP model. 32.
(33) BAP model. 33.
(34) BAP model. 34.
(35) Letter recognition knowing the scripter. 35.
(36) Letter recognition knowing the scripter. 36.
(37) Letter recognition knowing the scripter. 93,36%. 37.
(38) Scripter recognition knowing the letter. 79,5%. 38.
(39) Motor control. 39.
(40) Motor equivalence. 40.
(41) Motor equivalence. 41.
(42) Copy. Trace copy. Letter copy. 42.
(43) Letter recognition with motor simulation. 43.
(44) Letter recognition with motor simulation. 44.
(45) Letter recognition with motor simulation. 45.
(46) Perspectives (2) Back to speech perception and production. 46.
(47) Overview How to survive (perceive, reason, learn, decide and act) with incomplete information ? Probability as an alternative to logic How to develop better artifacts using Bayesian reasoning? Biological plausibility of Bayesian reasoning at a macroscopic level? Biological plausibility of Bayesian reasoning at a microscopic level?. 47.
(48) Amoeba How is it performing probabilistic inference?. Cell signaling 48.
(49) Microscopic Level Probabilistic inference by the biochemical mechanisms of phototransduction Activation of Phosphodiastrase Hydrolysis of cGMP Closure of Ca++ ion channels Decrease of Ca++ concentration Ca++ regulates Activation of GC GC catalyses production of cGMP. AMPLIFICATION Hyperpolarization of the Membrane.
(50) Want to know more?. Bayesian-Programming.org. Pierre.Bessiere@College-de-france.fr 50.
(51)
Documents relatifs
Based on this representation, Section 4 discusses how beliefs and back- ground knowledge influences the reasoning process and shows that the results can be modeled by computing
The force behind developing a two-level logic approach to reasoning about logic specifications is the necessity to treat both induction (and co-induction) and higher-order
qualitative modelling, cognitive maps, bayesian networks, fisher’s knowledge, fisheries management, qualitative decision support...
Here θ denotes a concatenation of the parameters defining the distribution of the profile of the seed cluster and φ denotes a concatenation of all the transition parameters of
To build on students’ intuitive and informal inferential reasoning, a Bayesian in- spired approach was adopted to design the Matching Tokens Game task in which students were asked to
However, it is important that the black artwork versions are used, as use of the Oxford blue artwork versions could result in a half-tone grey being produced.. Okd`rdmnsd9 do
The infrastructure we propose draws on both recent developments in universal logical reasoning in classical higher-order logic (HOL) [5] and the coalescence and improvements
function f (thin line); n=10 evaluations (squares) as obtained by the proposed algorithm using m = 800 for the Monte Carlo integration and Q = 20 for the tagged partition;
