Transcription
NEURAL NETWORKSTHE BRAIN OF THE TRULYAUTONOMOUS UAVStephen L. Thaler, Ph.D.President & CEO, Imagination Engines, Inc.40th Annual NDIA Air Targets, UAVs Range Operations Symposium 2002 Imagination Engines, Inc.
UNIVERSE IS A COMPLEX NETWORK 2002 Imagination Engines, Inc.
NEURAL NETS MODEL UNIVERSEConnection weights are tantamount to expansion coefficients within a curve fit.input pattern xcan be anybasis functionbecause of someinitiating event.because somethingelse happenedMLPs capturecomplex causalchains.something happenedoutput pattern FN(FN-1( F1(x) )) in contrast to the usual F(x) a 0F0(x) a2F1(x) aNFN(x) 2002 Imagination Engines, Inc.
MLPS* CAPTURE ENTITIES & CONNECTS* MLP Multi-Layer Perceptron, the workhorse of artificial neural networks.bitmaps of faces and other objects xclassificationlayereye detectorsEye, nose, mouth,ear, etc.detectingcolonies form.nose detectorsIf eyes, nose,mouth, and earsare detected thenface is present.constraint layer(s)BOOL isaFace(x)output patternsBOOL isaFoot(x) 2002 Imagination Engines, Inc.
CREATIVITY MACHINE PARADIGMhopping synapticperturbations“CM”inputs clamped for contextImaginationEngine(IE)Stream onof IdeasoutputsUS05659666, 08/19/199, Device for the Autonomous Generation of Useful Information 2002 Imagination Engines, Inc.
AUTONOMOUS WARHEAD ADAPTATION 2002 Imagination Engines, Inc.
AUTONOMOUS MOTION PLANNINGCreativity Machineimplemented cockroachsimulation 30 degree of freedom mastered in30 seconds. A 1,000 degree offreedom system represented inflight controls do not pose aproblem.GPS coordinates, chemical orbiological gradients, etc .User drags roach whilesimulation invents therequired legwork!All that need be supplied is a ‘will’ toproceed in a certain direction,toward a particular goal, asrepresented by the cursor drag. 2002 Imagination Engines, Inc.
SELF-TRAINING ANN OBJECT“STANNO”synapticadjustmentsinput patternsUntrainedNeural NetworkSynapticCorrectionsfeedbackANN-BasedNeural NetworkTrainerEvaluation ofErroroutputsUS05845271, 12/01/1998, Non-Algorithmically implemented artificial neural networks and components thereof 2002 Imagination Engines, Inc.
STANNO-BASED ZDATA GENERATOR 2002 Imagination Engines, Inc.
MEMBERSHIP / ANOMALY DETECTIONIntact feed forward passage through auto-associativenet, trained upon some interrelated group of patterns,signifies membership within that group of patterns.Features key togroup Gmembershipidentified inhidden layers.PRelationshipsbetweensensedfeatures testedagainst thoseof group G inoutput layer(s).GP’G {P1, P2, PN}P P’ P GP P’ P Gnovelty vector, δ P – P’ (x1, x2, x3, , xN)US05852816, 12/22/1998, Neural network based database scanning system 2002 Imagination Engines, Inc.
BATTLE DAMAGE ASSESSMENT 2002 Imagination Engines, Inc.
BATTLE DAMAGE ASSESSMENT 2002 Imagination Engines, Inc.
AUTONOMOUS TARGET RECOGNITION 2002 Imagination Engines, Inc.
AUTONOMOUS TARGETINGyquiescentxy(x, y)xyx(0,0) 2002 Imagination Engines, Inc.
AUTONOMOUS TARGETINGperturbedxy(x, y)(x’, y’)(x’’, y’’)x’y’x(0,0) 2002 Imagination Engines, Inc.
AUTONOMOUS TARGETINGyperturbedxyMembership(x, y)filter(x’, y’)δ(x’’, y’’)x’y’x(0,0) 2002 Imagination Engines, Inc.
AUTONOMOUS TARGETINGyperturbedxyMembership(x, y)filter(x’, y’)δ(x’’, y’’)x’y’x(0,0)US05852816, 12/22/1998, Neural network based database scanning system 2002 Imagination Engines, Inc.
AUTONOMOUS TARGETING 2002 Imagination Engines, Inc.
AUTONOMOUS TARGETINGObjective: Autonomously find the toy F4 attention windowrobot held camera 2002 Imagination Engines, Inc.
AUTONOMOUS, BRILLIANT UAVSCreativity Machines STANNOs Group/Anomaly FiltersAircraft / UAV Design . Let them design themselves.Web-based Logistical Support Let them search for their own components.Field Kit for Tailored Assembly . Let them recommend their own field configurations.Self-diagnosis . Let them tell us when / if they’re ready to go.Flight Control / Dogfight / Egress . Let them fly themselves.Massively Parallel Sensor Integration . Let them fuse inputs and resolve ambiguities.Mission / Sortie Planning Let them plan based on loosely posed objectives.Strategy / Tactics Let them improvise as battlefield evolves.Battle Damage Recovery . Let them repair themselves in flight.Low Observables Adaptation Let them reconfigure themselves to evade.Autonomous Targeting . Let them lock on without slow human judgment.Battle Damage Assessment . Let them evaluate their effects on target and react.Legal Repercussions . Let them be their own, instantaneous cyber-lawyer.Political/Philosophical Perception . Let them have similar motivating “feelings.” 2002 Imagination Engines, Inc.
2002 Imagination Engines, Inc. bitmaps of faces and other objects x output patterns classification layer constraint layer(s) Eye, nose, mouth, ear, etc.
