<?xml version="1.0" standalone="yes"?>
<Paper uid="C92-3147">
  <Title>B-SURE: A BELIEVED SITUATION AND UNCERTAIN-ACTION REPRESENTATION ENVIRONMENT</Title>
  <Section position="1" start_page="0" end_page="0" type="metho">
    <SectionTitle>
B-SURE: A BELIEVED SITUATION AND
UNCERTAIN-ACTION REPRESENTATION ENVIRONMENT
JOI/N K. MYI';1LS
ATR Interpreting Telephony Research l,aboratories
</SectionTitle>
    <Paragraph position="0"/>
  </Section>
  <Section position="2" start_page="0" end_page="0" type="metho">
    <SectionTitle>
Abstract
</SectionTitle>
    <Paragraph position="0"> This paper l)reseltts a system that is c~q)abte of representing situs,loss, states, anci nondeterminlstic lIOIlinOllOtollic Oil'COllie actions ,lccmTillg ill multiple possible worlds. :\['he systcnl sup ports explicit representations of actions and situations used in intentional action theory and situation theory. \[lath types mid instances ere supported. Situations C/md statc&amp;quot;a before a,d aftel-llOllll/OllOtOnic actions c~ul be repl'esellted shnultaneously.</Paragraph>
    <Paragraph position="1"> Agents have free will as to whether to choose to peiform an ac:lion or not. Situations itud actions can have expected values, allowing the system to support decision making anti decision-based pleal isfferencing. The system cau perform global reasoning simultaneously across multiple possible worlds, without being forced to extend each world explicitly. The resulting system is useful for retch *tatural language t~-~ks a~ plan recognition, intentions modeling, attd parallel ta.~k scheduling.</Paragraph>
  </Section>
  <Section position="3" start_page="0" end_page="0" type="metho">
    <SectionTitle>
1. Introduction
</SectionTitle>
    <Paragraph position="0"> The key to good reasoning is a powerful representation system that is able to accuratcly model details of a problem. Once a good represent.at,on has been established, problem computations often become straightforward.</Paragraph>
    <Paragraph position="1"> IC/~ecent advances in situation theory \[BP83,Bax89\] and the theory of intentions \[Bra87\] have offered ninny new insights on significant problems found in natural-language understanding. However, these thearies offer philosophical approaches only, and do not give instructions for building concrete reprcsentation and reasoning engines.</Paragraph>
    <Paragraph position="2"> At the same time, the software systems that have been built for reasoning and representation fall short in any ruunber of areas. Production systems and semantic networks can follow chains of inferences, but can only represent one possible world at a time they cannot reason with states that are both possibly true and possibly not true, while keeping the chains of resulting inferences separatc. Most plaslnets work with limited possible worlds, but callnot reason and perform inferences across multiple worlds at the salne time. The classical ATMS 1 call represent and reason with multiple timeless possible worlds, but calmot represent actions \[dK86\]-in particular, non-monotonic actions where a retracted state is both believed'to be truc in the world before ttte action takes place, and believed to be not true in the world representing the situation after the retracting action has taken place, camtot be represented. In addition, the ATMS only represents propositions that are instant'steal constants or Skolem constants; it does not represent uninstantiated variables. A modified ATMS that can represent nonmonotonic transitions between worlds has been developed \[MN86\], but this system does not explicitly represent situation types and instances, action events, nor nondeterminism. Most plan inference systems have ignored free will and the 1Aanumption-13tmed Truth Maintenta*ce Systenl \[dK86\] explicit representation of the right to choose actions, e.g. to choose to he uncooperative. Almost all prcvious systems |lave ignored the nondetermimstie quality of real-world actions that aecessitatcs commitmeat ill intentions. Real actions call result in one of several possible outcomc situations, where,xs alolost all previous systems are are completely unable to model uondcterminlstie outcomes. Only dccisionanalysis systems have modeled cxpected wdues of actions, alnl they do not support inferencing. See \[BL85\] fi)r an excellent summary of issues.</Paragraph>
    <Paragraph position="3"> Tile B-SURI~ (Believed Situation and Uncertainaction Representation Environment) packagc is an implemented system that supports representation, phmnmg, decision-making, and,plan recognition using probabilistic ,and uncertain actions with non detcrministic outcomes in multiple possiblc action worlds. Situations, states, and action events ,axe all represented explicitly, using types (wtriables) and instances. The B-SURE systcm is iml)lemented as a series of extensions to a classical ATMS. The resulting system is very useful, and is being used in plan recognition, intentional agent, and scheduling research. 2. Situation Theory In \[BP83\], situations are divided into the categories abstract and real, and also into the categories &amp;quot;states of affairs&amp;quot; alid &amp;quot;courses of events&amp;quot;. Abstract situations denote situations that are mental representations. All the situations discussed in this paper are &amp;quot;abstract situations&amp;quot;. Real sitnations denote situations as they actually are in the real world. Since it basically never makes sense to talk about real situations in tile computer, there is no need to snpply these in a representation environment. &amp;quot;States of affairs&amp;quot; correspond to situations that axe static, called simply situations in this paper. &amp;quot;Courses of events&amp;quot; correspond to situations that describe actions that are being executed, called action events or actions in this paper. Barwise and Perry also make use of &amp;quot;relations&amp;quot; defined over '*individuals&amp;quot; and &amp;quot;space-time locations&amp;quot;. This paper takes as primitive the expression of a relation, which will be termed a stats. The user is free to mention individuals or space-time locations in state descriptions as desired. State descriptions may be represented nsing logical forms, feature structures, or other methods~sinee the contents of states are not used by B-SURE except for output, it does not matter. States, situations, and actions axe assigned one of the belief values {definitely believed tzate, possibly believed true, not believed true, believed not true, not believed}, otherwise known ms {actua/, possible, bypotheticed..</Paragraph>
    <Paragraph position="4"> inconsistent, null}, corresponding to the amount of snpport off, red by tile system's underlying ATMS  One model of intentions states that an intention is a choice to perform an action, plus a commitment to obtaining its desired outcome\[CL87\]. With deterministic action outcomes, there is no real need for endeavoring \[Brag7\], since once the action has been started, it is guaranteed to finish properly. Many planners in fact operate in this &amp;quot;fire and forget&amp;quot; mode. However, once it is acknowledged that action execution is in fact nondeterministic and can have undesirable outcomes, the need for endeavoring becomes clear. The planner must predict the likelihood of possible outcomes happening, and judge which action sequence offers the best chances. It must interactively maintain a history of past endeavors and results, and modify its future behavior based on current outcomes. Acting intentionally becomes significantly more interesting and realistic with the explicit representation of possible chains of nondeterministic actions.</Paragraph>
  </Section>
  <Section position="4" start_page="0" end_page="0" type="metho">
    <SectionTitle>
4. Previous Efforts
</SectionTitle>
    <Paragraph position="0"> DeKleer \[dK86\] presents the first ATMS. Morris and Nado \[MN86\] present an ATMS that can represent nonmonotonic transitions, but do not handle probabilities, uncertainties, explicit situation types, state types, nor action events. Tile research of Allen (e.g.</Paragraph>
    <Paragraph position="1"> \[AK83, A1187\]), who uses a predicate-calculus representation, offers some of the best multiple-worlds (deterministic) action representation in this field.</Paragraph>
    <Paragraph position="2"> Charniak and Goldman \[CG89\] use probabilities and Bayesian nets to represcnt the truth value of probabilistic statements and attack story understanding.</Paragraph>
    <Paragraph position="3"> Although nondetcrministic-outcomc actions are not represented, and Bayesian nets cannot support global inferencing with nonnronotonic actions, their work is important. Norvig and Wilensky \[NW90\] comment on problems of probabilistic statements. &amp;quot;1'he most similar work is recent research by Rao and Georgeff (e.g. \[RGgl\]), who use a modal logic instead of an ATMS to represent nondeterministic actions.</Paragraph>
  </Section>
  <Section position="5" start_page="0" end_page="0" type="metho">
    <SectionTitle>
5. B-SURE Entities &amp; Implementation
</SectionTitle>
    <Paragraph position="0"> The underlying ATMS works with nodes, assnmptions, and implications (justifications). See \[dK86\]. A slate consists of a proposition about the world.</Paragraph>
    <Paragraph position="1"> States are primitives. A situation is a set of positive and negative (withdrawn) states. An action event represents the state that &amp;quot;execution of the action has started&amp;quot;. States, situations, and actions have types and instances. See figure 1. (The abridged representation of figure 1 is shown in figure 2.) Existance of an instance in a world always implies existance of its type. A chooses node is an assumption associated with an action instance that represents whether an agent chooses to execute that action or not. The chooses assumption together with the starting situation instance imply the action instance. Since an agent typically can only execute one action in a given situation, the situation's ensuing chooses assumptions are rendered mutually exclusive (pairwise &amp;quot;nogood&amp;quot;). Action types have precondition situation types. Action instances are instantiated from types by first verifying that the precondition situation type is believed true in that world. Action instances transition from a starting situation instance to ouc of a number of known nondeterministic outcome situation instances.</Paragraph>
    <Paragraph position="2"> Actions have transitions. A transition has an outcon,e situation and a probability or an uncertainty.</Paragraph>
    <Paragraph position="3"> An uncerlaiuty is defined as a probability random variable of range I0, 1\] together with an associated second-order probability distrilmtion. Uncertainties are initialized using maximum-entropy theory, and get updated as outcome observations are taken, to enable the system to learn and estimate possible probabilities. See Section 6. Uncertainties are used to represent confidence in likelihood values and to make decisions regarding information-gathering activity. The calculus of uncertainties is too complex to explore further here, and is not required for understanding tile mum capabilities of the representation; probabilities are sufficient. Transitions can be types or instances.</Paragraph>
    <Paragraph position="4"> ACRES DE COLING-92, NANIT~. 23-28 AOt~rr 1992 9 6 2 PROC. OF COL1NG-92, NANTES, AUO. 23-28. 1992 A transition instance is defined as a happens mssumption. An action instance, together with a happens assumption, iulply the corresponding outcome situation instance. Typically only one outcome situation can occur from a giveu action instance, so the action's happens assun-|ptions are nlade mutually exclusive. A situation type is implied by its state types. When an outcome situation instance is iastantiated, all of its new positive states are instantiated and all of its ohl negative states are retracted. A positive nonpermanent state instance is implied by a nol-relracled-yel assumption. The outcome situation instance remembers these. Situation and action instances store an explicit environment history of all added state, chooses, and happens assumptions that are currently believed true in that possible world's timelinc. A negative state is retracted by making the situation instance and the state's &amp;quot;not-rctrasted yet&amp;quot; ,assumption mutually inconsistent, and deleting the state's assumption from the outcome situation's environment history. A state type or instance or situation type's belief value in a particular world is found by testing that node against a situation instance's environment history.</Paragraph>
    <Paragraph position="5"> Situation types and instances can have values. Actions can have costs. The expected value of an action is determined by summing the transition probabilities times the expected values of the outcome situations, when known, and subtracting its cost. Tim expected value of a nonvalued situation instance is determined by maximizing the expected values of the possible subsequent actions, when known. In this manner, decision theory determines the course of action with the maximum expected value at any one situation, for a planning agent. This can bc used to predict the probable next course of action of a planning agent by an observing agent performing plan recognition (actually, &amp;quot;decision recognition&amp;quot;) \[Mye91\].</Paragraph>
  </Section>
class="xml-element"></Paper>