Opmaker2: Efficient Action Schema Acquisition T.L.McCluskey, S.N.Cresswell, N. E. Richardson and M.M.West The University of Huddersfield,UK

Slides:



Advertisements
Similar presentations
Cover Slide ~ Semantic Web Pilot Program Elements ~ Presented by Cola Atkinson – BBN 07 May 2003.
Advertisements

Knowledge Engineering for Planning Domain Design Ron Simpson University of Huddersfield.
ARCHITECTURES FOR ARTIFICIAL INTELLIGENCE SYSTEMS
AI – Week 17 Machine Learning Applied to AI Planning: LOCM Lee McCluskey, room 2/09
1 Requirements and the Software Lifecycle The traditional software process models Waterfall model Spiral model The iterative approach Chapter 3.
CHAPTER 1 SOFTWARE DEVELOPMENT. 2 Goals of software development Aspects of software quality Development life cycle models Basic concepts of algorithm.
Planning and Scheduling. 2 USC INFORMATION SCIENCES INSTITUTE Some background Many planning problems have a time-dependent component –  actions happen.
AI – Week 23 – TERM 2 Machine Learning and Natural Language Processing Lee McCluskey, room 3/10
Ontologies and the Semantic Web by Ian Horrocks presented by Thomas Packer 1.
Planning Copyright, 1996 © Dale Carnegie & Associates, Inc. Chapter 11.
Towards An Algebraic Formulation of Domain Definitions using Parameterised Machines T. L. McCluskey and R. M.Simpson School of Computing and Engineering.
School of Computing and Mathematics, University of Huddersfield The Induction of Operator Descriptions from Examples and Structural Domain Knowledge Lee.
Chair of Software Engineering Einführung in die Programmierung Introduction to Programming Prof. Dr. Bertrand Meyer Lecture 17: Topological Sort Algorithm.
GIPO II: HTN Planning in a Tool- supported Knowledge Engineering Environment Lee McCluskey Donghong Liu Ron Simpson Department of Computing and Mathematical.
Marakas: Decision Support Systems, 2nd Edition © 2003, Prentice-Hall Chapter Chapter 4: Modeling Decision Processes Decision Support Systems in the.
Knowledge and Systems Research Group, University of Huddersfield B vs OCL: Comparing Specification Languages for Planning Domains Diane Kitchin, Lee McCluskey,
Knowledge Acquisitioning. Definition The transfer and transformation of potential problem solving expertise from some knowledge source to a program.
Brent Dingle Marco A. Morales Texas A&M University, Spring 2002
Lee McCluskey, University of Huddersfield - EKAW'04 Knowledge Formulation for AI Planning Lee McCluskey Ron Simpson Artform research group Department of.
School of Computing and Mathematics, University of Huddersfield An Interactive Method for Inducing Operator Descriptions Lee McCluskey Beth Richardson.
Learning from Observations Copyright, 1996 © Dale Carnegie & Associates, Inc. Chapter 18 Fall 2004.
School of Computing and Mathematics, University of Huddersfield Knowledge Engineering: Issues for the Planning Community Lee McCluskey Department of Computing.
Creating Architectural Descriptions. Outline Standardizing architectural descriptions: The IEEE has published, “Recommended Practice for Architectural.
Planning Copyright, 1996 © Dale Carnegie & Associates, Inc. Chapter 11.
1 State-Based Testing of Ajax Web Applications A. Marchetto, P. Tonella and F. Ricca CMSC737 Spring 2008 Shashvat A Thakor.
School of Computing and Mathematics, University of Huddersfield PDDL and other languages.. Lee McCluskey Department of Computing and Mathematical Sciences,
School of Computing and Mathematics, University of Huddersfield Week 21: Knowledge Acquisition / GIPO Lee McCluskey, room 2/09
Object-oriented design CS 345 September 20,2002. Unavoidable Complexity Many software systems are very complex: –Many developers –Ongoing lifespan –Large.
Domain-Specific Software Engineering Alex Adamec.
Object-Oriented Analysis and Design
Reasoning with context in the Semantic Web … or contextualizing ontologies Fausto Giunchiglia July 23, 2004.
An Introduction to MBT  what, why and when 张 坚
Process Flowsheet Generation & Design Through a Group Contribution Approach Lo ï c d ’ Anterroches CAPEC Friday Morning Seminar, Spring 2005.
1 Intelligent Systems ISCRAM 2013 Validating Procedural Knowledge in the Open Virtual Collaboration Environment Gerhard Wickler AIAI, University.
Chapter 6 Supplement Knowledge Engineering and Acquisition Chapter 6 Supplement.
Chapter 2 소프트웨어공학 Software Engineering 임현승 강원대학교
 Knowledge Acquisition  Machine Learning. The transfer and transformation of potential problem solving expertise from some knowledge source to a program.
Aquinas Hobor and Cristian Gherghina (National University of Singapore) TexPoint fonts used in EMF. Read the TexPoint manual before you delete this box.:
CHA2555 Week2: Knowledge Representation, Logic and Planning Lee McCluskey First term:
Validated Model Transformation Tihamér Levendovszky Budapest University of Technology and Economics Department of Automation and Applied Informatics Applied.
The european ITM Task Force data structure F. Imbeaux.
Implementing a Domain-Specific Modeling Environment For a Family of Thick-Client GUI Components Milosz Muszynski Tanner AG
Introduction to Problem Solving. Steps in Programming A Very Simplified Picture –Problem Definition & Analysis – High Level Strategy for a solution –Arriving.
Artificial intelligence methods in the CO 2 permission market simulation Jarosław Stańczak *, Piotr Pałka **, Zbigniew Nahorski * * Systems Research Institute,
A Universe-Type-Based Verification Technique for Mutable Static Fields and Methods Alexander J Summers Sophia Drossopoulou Imperial College London Peter.
Simultaneously Learning and Filtering Juan F. Mancilla-Caceres CS498EA - Fall 2011 Some slides from Connecting Learning and Logic, Eyal Amir 2006.
Bing LiuCS Department, UIC1 Chapter 8: Semi-supervised learning.
Data Structures and Algorithms Dr. Tehseen Zia Assistant Professor Dept. Computer Science and IT University of Sargodha Lecture 1.
MODEL-BASED SOFTWARE ARCHITECTURES.  Models of software are used in an increasing number of projects to handle the complexity of application domains.
1 Analysing system-user cooperation in KADS H. P. de Greef and J. A. Breuker, Department of Social Science Informatics, University of Amsterdam Knowledge.
AI Lecture 17 Planning Noémie Elhadad (substituting for Prof. McKeown)
Issues in Ontology-based Information integration By Zhan Cui, Dean Jones and Paul O’Brien.
Scientific Debugging. Errors in Software Errors are unexpected behaviors or outputs in programs As long as software is developed by humans, it will contain.
Rational Agency CSMC Introduction to Artificial Intelligence January 8, 2007.
Course Overview  What is AI?  What are the Major Challenges?  What are the Main Techniques?  Where are we failing, and why?  Step back and look at.
A Roadmap towards Machine Intelligence
Finite State Machines (FSM) OR Finite State Automation (FSA) - are models of the behaviors of a system or a complex object, with a limited number of defined.
A System for Automatic Personalized Tracking of Scientific Literature on the Web Tzachi Perlstein Yael Nir.
A Portrait of the Semantic Web in Action Jeff Heflin and James Hendler IEEE Intelligent Systems December 6, 2010 Hyewon Lim.
Natural Language Processing AI Revision Lee McCluskey, room 2/07
Of An Expert System.  Introduction  What is AI?  Intelligent in Human & Machine? What is Expert System? How are Expert System used? Elements of ES.
AI – Week 16 Machine Learning Applied to AI Planning Lee McCluskey, room 2/09
Concept mining for programming automation. Problem ➲ A lot of trivial tasks that could be automated – Add field Patronim on Customer page. – Remove field.
Michael J. Voss and Rudolf Eigenmann PPoPP, ‘01 (Presented by Kanad Sinha)
 Knowledge Acquisition  Machine Learning. The transfer and transformation of potential problem solving expertise from some knowledge source to a program.
Biointelligence Lab School of Computer Sci. & Eng. Seoul National University Artificial Intelligence Chapter 8 Uninformed Search.
Done Done Course Overview What is AI? What are the Major Challenges?
Parallel Programming By J. H. Wang May 2, 2017.
Software Project Planning &
Planning CSE 573 A handful of GENERAL SEARCH TECHNIQUES lie at the heart of practically all work in AI We will encounter the SAME PRINCIPLES again and.
Presentation transcript:

Opmaker2: Efficient Action Schema Acquisition T.L.McCluskey, S.N.Cresswell, N. E. Richardson and M.M.West The University of Huddersfield,UK

Knowledge Engineering and Intelligent Interfaces Motivation – Intelligent Agents that can Plan? The availability of knowledge via ontologies is growing n it is not unreasonable that an agent will be able to acquire and refine such knowledge with some degree of autonomy. n A necessary precondition of the use of (current) automated planning technology is that there exists detailed knowledge: action specifications and heuristics n the amount of effort needed to encode bug free, accurate action specifications and planning heuristics, and to maintain them, is significant. for every agent that can perform planning, must we hand code and hand maintain its action descriptions and heuristics? No, if agents are to achieve this kind of autonomy, then they should be capable of learning and refining action knowledge and heuristics.

Knowledge Engineering and Intelligent Interfaces Vision Agent Application Ontologies Plan Engines Learning phase Agent capable of deliberative planning

Knowledge Engineering and Intelligent Interfaces Planning and Machine Learning – moved on? Eg my work in 1987 (IJCAI-87, IWML-87) deriving planning heuristics (macros, operator choice, goal- orders) offline YES – must derive heuristics + dynamic domain model

Knowledge Engineering and Intelligent Interfaces General Idea of Opmaker2 A knowledge base for planning can be divided into 3: - the easy (static) part of a domain model - The (hard) dynamic part - Planning heuristics The idea of Opmaker is to learn the (hard) dynamic part and the planning heuristics from “training sequences” within the context of an existing (static) part of a domain model

Knowledge Engineering and Intelligent Interfaces Learning Problem - INPUT Input: (a) knowledge of objects, and collections of similar objects making up distinct classes, possible states of a typical object of each class, and state invariants (b) Knowledge of existing plans that other agents, or a trainer, has used [in terms of verbs and affected objects] with initial state and goal state for affected objects the agent does not have an explicit specification of actions in such a way that it can reason about their synthesis (or the agent does have such a specification but needs to refine, maintain or evolve it).

Knowledge Engineering and Intelligent Interfaces Example of (b) Example training sequence from “Extended Tyre Domain” 1. name: do_up; unchanging: wrench0,jack0, trim1; changing: hub1,nuts1 2. name: jack_down; unchanging:: nil changing: hub1,jack0 3. name: tighten; unchanging: wrench0,hub1,trim1; changing: nuts1 4. name: apply_trim; unchanging: hub1; changing: trim1,wheel5

Knowledge Engineering and Intelligent Interfaces Learning Problem - OUTPUT 1. A full parameterised specification of actions which can be used to do planning 2. Planning heuristics. 3. Refinement of any existing parameterised specification of actions, and heuristics, that it currently holds. Results: Opmaker2 outputs 1, plus 2 in the form of HTN methods. Trying to avoid 3.

Knowledge Engineering and Intelligent Interfaces Opmaker vs Opmaker2 Opmaker2 INPUT: Partial (static) Domain Theory Training sequences (.. extra Invariants) OUTPUT: Primitive Operator Schema HTN Schema Opmaker (in GIPO) INPUT: Partial (static) Domain Theory Training sequences Intermediate State Details OUTPUT: Primitive Operator Schema But hasn’t this been done before – by Opmaker(1) (2002) for example? Painful part!

Knowledge Engineering and Intelligent Interfaces Method Overview (1).use a set of heuristics and inferences to track the changing states (‘object trace’) of each changing object referred to within a training example; produce a set of candidate operators (2). use domain invariants to reduce the set of candidate operators (to a singleton) (3). use the techniques of the original Opmaker algorithm to generalise object references and create parameterised operator schema from the specific object transitions extracted in (i) from the training examples.

Knowledge Engineering and Intelligent Interfaces Method Overview for (1) Track an object O in a training sequence Id1,Id2,…,Idk,…, Idn: Where Idk includes the last mention of changing O O.i O.f Method seeks to fill in state transition details.. KNOWN UNKNOWN Id1 Idk

Knowledge Engineering and Intelligent Interfaces Method Overview for (1) Track an object O in a training sequence Id1,Id2,…,Idk,…, Idn: Where Idk includes the last mention of changing O O.i O.f..by finding the Object’s trace including any associations it makes Idk Id1

Knowledge Engineering and Intelligent Interfaces Method Overview for (1) Considering a changing object O at training action with id ‘Idx’.. If Idx is not the last change of O then we potentially have a choice of target state for O … O current state = O.cIdx Same state State that has parameter of object type NOT in Idx Potential new states

Knowledge Engineering and Intelligent Interfaces Back to Example id1: do_up; unchanging: wrench0,jack0, trim1; changing: hub1,nuts1 id2: jack_down; unchanging:: nil changing: hub1,jack0 id3: tighten; unchanging: wrench0,hub1,trim1; changing: nuts1 Id4: apply_trim; unchanging: hub1; changing: trim1,wheel5

Knowledge Engineering and Intelligent Interfaces Method Overview for (1) - example Considering a changing object hub1 at training action ‘do_up’.. If do_up is not the last change of hub1 then we potentially have a choice of target state for hub1 … hub1 current state = unfastened(hub1), jacked_up(hub1,jack0) do_up Same state Potential new states (see PAPER for definitions) S1 S2S3S4

Knowledge Engineering and Intelligent Interfaces Method Overview – result of step (1) Form sets of candidate operators for each Id, of form - Id - x - Set of parameters - Set of prevail conditions (unchanging objects) - Set of object transitions

Knowledge Engineering and Intelligent Interfaces Method Overview – step (2) use domain invariants to reduce the set of candidate operators (to a singleton) Eliminate a candidate that is logically consistent with domain invariants

Knowledge Engineering and Intelligent Interfaces Method Overview – step (3) use the techniques of the original Opmaker algorithm to generalise object references and create parameterised operator schema from the specific object transitions extracted in (i) from the training examples. Generalise those left using OpMaker to create operator schema

Knowledge Engineering and Intelligent Interfaces Producing Heuristics - Our earlier paper (2002) on Opmaker indicated how the training sequences can be made into ‘canned plans’ and used as components in planning using the Hybrid planner HyHTN - Even earlier work (1997) showed how invariants are key to producing goal ordering heuristics - In Opmaker2 we have implemented the former method and the canned plans are produced with the primitive operator schema

Knowledge Engineering and Intelligent Interfaces Experiments 3 Domains considered : Extended Tyre Domain, Hiking Domain, Blocks Domain Enough training sequences to include instances of all required operator schema Experimental Criteria: - A. ‘How many’ axioms have to be added to the static domain model to make sure the operator set is unique? - B. Are the operator sets induced capable of being used in solving planning problems? - C. With the canned plans introduced, is plan generation time reduced? - D. Does the performance of the learning program vary a great deal with different training sequences

Knowledge Engineering and Intelligent Interfaces Experimental Results A. ‘How many’ axioms have to be added to the static domain model to make sure the operator set is unique? - Tyre Domain - 8, Hiking Domain - 0, Blocks Domain - 4 B. Are the operator sets induced capable of being used in solving planning problems? Yes, in all 3 cases C. With the canned plans introduced, is plan generation time reduced? - Tyre Domain, Hiking Domain, Blocks Domain Yes, in all 3 cases (no graphs to show as yet..) D. Still much to do BUT LONG training sequences do not seem to cause a problem – we tried a 22 element solution in the Blocks World and it produced a single set of operator schema BUT result with blocks domain threw up some problems because of the multiple object references of the same class within one operator schema

Knowledge Engineering and Intelligent Interfaces Conclusions and Future Work Essentially, we have traded off effort to produce a detailed “static” domain model vs effort to produce operator schema modelling actions because (a) Static knowledge is more easily available (b) Our experience of operator building in GIPO II showed that creating the parameterised specs to be the hardest part of domain encoding Open Question for future work: How to induce operator schema which require constraints mot appearing in the input object state definitions How to guarantee that a set of invariants is strong enough to deliver a working set of operator schema?