<?xml version="1.0" standalone="yes"?>
<Paper uid="P98-1084">
  <Title>Integrating Text Plans for Conciseness and Coherence*</Title>
  <Section position="4" start_page="512" end_page="512" type="metho">
    <SectionTitle>
3 Other Work
</SectionTitle>
    <Paragraph position="0"> The idea of domain-independent text planning rules is not new. Appelt (1985) used &amp;quot;interactions typical of linguistic actions&amp;quot; to design critics for action subsumption in KAMP. REVISOR (Callaway and Lester, 1997) used domain-independent operators for revision of a text plan for explanation. Because our rules operate on full RST-style text plans that include communicative goals, the rules can be designed to integrate the text plans in ways that still satisfy those goals.</Paragraph>
    <Paragraph position="1"> The Sentence Planner (Wanner and Hovy, 1996) uses rules to refine a single initial tree representation. In contrast, RTPI operates on sets of complete, independent text plan trees. And while REVISOR handles clause aggregation, and Sentence Planner removes redundancies by aggregating neighboring expressions, neither of them addresses the aggregation of communicative goals (often requiring reorganization), the TraumaTIQ critiques: Caution: check for medication allergies.and do a laparotomy immediately to treat the intra-abdominal injury.</Paragraph>
    <Paragraph position="2"> Consider checking for medication allergies now to treat a possible GI tract injury.</Paragraph>
    <Paragraph position="3"> Please remember to check .for medication allergies before you give antibiotics.</Paragraph>
    <Paragraph position="4"> Message from RTPI integrated plan: Caution: check for medication allergies to treat the intra-abdominal injury and a possible GI tract injury, and do it before giving antibiotics. Then do a laparotomy to complete treating the intra-abdominal injury.</Paragraph>
    <Paragraph position="5">  revision and integration of text plans to remove conflict, or the exploiting of relations between communicative goals as done by RTPI. Similarly, WISHFUL (Zukerman and McConachy, 1995) includes an optimization phase during which it chooses the optimal way to achieve a set of related communicative goals. However, the system can choose to eliminate propositions and does not have to deal with potential conflict within the information to be conveyed.</Paragraph>
  </Section>
  <Section position="5" start_page="512" end_page="513" type="metho">
    <SectionTitle>
4 System Parameters
</SectionTitle>
    <Paragraph position="0"> Although RTPI's rules are intended to be domain-independent, environmental factors such as the purpose of the messages and the social role of the system affect how individual text plans should be integrated. For example, if the system's purpose is to provide directions for performing a task, then an ordered set of actions will be acceptable; in contrast, if the system's purpose is decision support, with the user retaining responsibility for the selected actions, then a better organization will be one in which actions are grouped in terms of the objectives they achieve (see Section 5.1.1).</Paragraph>
    <Paragraph position="1"> Similarly, in some environments it might be reasonable to resolve conflict by omitting communicative goals that conflict with the system's action recommendations, while in other environments such omission is undesirable (see Section 5.1.2).</Paragraph>
    <Paragraph position="2"> RTPI has a set of system parameters that capture these environmental factors. These parameters affect what rules are applied, and in some cases how they are applied. They allow characteristics of the output text plans to be tailored to broad classes of domains, giving the system the flexibility to be effective over a wide range of problems.</Paragraph>
  </Section>
  <Section position="6" start_page="513" end_page="514" type="metho">
    <SectionTitle>
5 The Rule-Base
</SectionTitle>
    <Paragraph position="0"> RTPTs input consists of a set of text plans, each of which has a top-level communicative goal. Rhetorical Structure Theory (Mann and Thompson, 1987) posits that a coherent text plan consists of segments related to one another by rhetorical relations such as MOTIVA-TION or BACKGROUND. Each text plan presented to RTPI is a tree structure in which individual nodes are related by RST-style relations. The top-level communicative goal for each text plan is expressed as an intended effect on the user's mental state (Moore, 1995), such as (GOAL USER (DO ACTION27)), The kinds of goals that RTPI handles are typical of critiquing systems, systems that provide instructions for performing a task, etc. These goals may consist of getting the user to perform actions, refrain from performing actions, use an alternate method to achieve a goal, or recognize the temporal constraints on actions.</Paragraph>
    <Paragraph position="1"> Rules are defined in terms of tree specifications and operators, and are stylistically similar to the kinds of rules proposed in (Wanner and Hovy, 1996). When all the tree specifications are matched, the score function of the rule is evaluated. The score function is a heuristic specific to each rule, and is used to determine which rule instantiation has the best potential text realization. Scores for aggregation rules, for example, measure the opportunity to reduce repetition through aggregation, subsumption, or pronominal reference, and penalize for paragraph complexity.</Paragraph>
    <Paragraph position="2"> Once a rule instantiation is chosen, the system performs any substitutions, pruning, and moving of branches specified by the rule's operators. The rules currently in use operate on text plan trees in a pairwise fashion, and recursively add more text plans to larger, already integrated plans.</Paragraph>
    <Section position="1" start_page="513" end_page="514" type="sub_section">
      <SectionTitle>
5.1 Classes of Rules
</SectionTitle>
      <Paragraph position="0"> RTPI has three classes of rules, all of which produce an integrated text plan from separate text plans. The classes of rules correlate with the three categories of problems that we identified from our analysis of TraumaTIQ's critiques, namely, the need to: 1) aggregate communicative goals to achieve more succinct text plans; 2) resolve conflict among text plans; and 3) exploit the relationships between communicative goals to enhance coherence.</Paragraph>
      <Paragraph position="1">  Our analysis of TraumaTIQ's output showed that one prevalent problem was informational overlap, i.e. the same actions and objectives often appeared as part of several different input text plans, and thus the resulting messages</Paragraph>
      <Paragraph position="3"> appear repetitious. Aggregation of the communicative goals associated with these actions and objectives allows RTPI to make the message more concise.</Paragraph>
      <Paragraph position="4"> Aggregation of overlapping communicative goals is not usually straightforward, however, and often requires substantial reorganizing of the trees. Our approach was to draw on the ordered, multi-nuclear SEQUENCE relation of RST. We posited that separate plans with overlapping communicative goals could often be reorganized as a sequence of communicative goals in a single plan. The recommended actions can be distributed over the sequentially related goals as long as the new plan captures the relationships between the actions and their motivations given in the original plans.</Paragraph>
      <Paragraph position="5"> For example, one complex class of aggregation is the integration of text plans that have overlapping actions or objectives, but also contain actions and objectives that do not overlap. When those that overlap can be placed together as part of a valid sequence, a multi-part message can be generated. RTPI produces an integrated text plan comprised of sequentially related segments, with the middle segment conveying the shared actions and their collected motivations.</Paragraph>
      <Paragraph position="6"> The other segments convey the actions that temporally precede or follow the shared actions, and are also presented with their motivations.</Paragraph>
      <Paragraph position="7"> For example (Fig. 5), suppose that one text plan has the goal of getting the user to perform actions A0, A2, and A3 to achieve G1, while a second text plan has a goal of getting the user to perform A1,A2, A3, and A4 to achieve G2.</Paragraph>
      <Paragraph position="8"> Figure 3 presents the text plan resulting from the application of this rule. Realization of this text plan in English produces the message: Do AO as part of G1, and A1 as part of G2.</Paragraph>
      <Paragraph position="9"> Next do A2 and A3 to address both of these goals. Then do A4 to complete G2.</Paragraph>
      <Paragraph position="10">  This kind of aggregation is especially appropriate in a domain (such as trauma care) where the clause re-ordering normally applied to enable aggregation (e.g. Sentence Planner) is restricted by the partial ordering of sequenced instructions. null RTPI can also handle aggregation when actions or objectives are shared between different kinds of communicative goals. The bottom part of Figure 1 is the text realized from a text plan that was produced by the application of two rules to three initial text plans: one rule that applies to trees of the same form, and one that applies to two distinct forms. The first rule aggregates the communicative goal</Paragraph>
    </Section>
  </Section>
  <Section position="7" start_page="514" end_page="516" type="metho">
    <SectionTitle>
(GOAL USER (DO USER check_med_allergies)) that exists
</SectionTitle>
    <Paragraph position="0"> in two of the text plans. The second rule looks for overlap between the communicative goal of getting the user to do an action and the goal of having the user recognize a temporal constraint on actions. The application of these two rules to the text plans of the three initial messages shown in the top part of Figure 1 creates the integrated text plan shown in Figure 4 whose English realization appears in the bottom part of Figure 1.</Paragraph>
    <Paragraph position="1"> RTPI's parameter settings capture aspects of the environment in which the messages will be generated that will affect the kind of aggregation that is most appropriate. The settings for aggregation determine whether RTPI emphasizes actions or objectives. In the latter case (appropriate in the trauma decision-support environment), an arbitrary limit of three is placed on the number of sequentially related segments in a multi-part message, though each segment can still address multiple goals. This allows the reorganization of communicative goals to enable aggregation while maintaining focus on objectives. null  The ability to recognize and resolve conflict is required in a text planner because both the appearance and resolution of conflict can be the result of text structure. RTPI identifies and resolves a class of domain-independent conflict, with the resolution strategies dependent upon the social relationship between the user and the system. In addition, the system allows the user to add rules for domain-specific classes of conflict. null One class of conflict that can best be resolved at the text planning level results from implicit messages in text. Resolving conflict of this kind within independent modules of a critiquing system would require sharing extensive knowledge, thereby violating modularity concepts and making the planning process much more complex.</Paragraph>
    <Paragraph position="2"> For example, suppose that the user has conveyed an intention to achieve a particular objective by performing act Au. One system module might post the communicative goal of getting the user to recognize that act Ap must precede Au, while a different module posts the goal of getting the user to achieve the objective by executing As instead of Au. While each of these communicative goals might be well-motivated and coherent in isolation, together they are incoherent, since the first presumes that Au will be executed, while the second recommends retracting the intention to perform Au. A text planner with access to both of these top-level communicative goals and their text plans can recognize this implicit conflict and revise and integrate the text plans to resolve it.</Paragraph>
    <Paragraph position="3"> There are many ways to unambiguously resolve this class of implicit conflict. Strategy selection depends on the social relationship between the system and the user, as captured by three of RTPTs parameter settings. This relationship is defined by the relative levels of knowledge, expertise, and responsibility of the system and user. Three strategies used by our system, and their motivations, are: I. Discard communicative goals that implicitly conflict with a system recommendation. In the above example, this would result in a text plan that only recommends doing As  with top level goals of the form (GOAL USER (DO ...))uses two trees from Fig. 5 to make a tree with the two subtrees labelled (1) and (2). Next, a rule that places scheduling trees ( (GOAL U (KNOW</Paragraph>
    <Paragraph position="5"> tree. A domain specific realizer traverses the tree and inserts cue words and conjunctions based on relations.</Paragraph>
    <Paragraph position="6"> instead of An. This strategy would be appropriate if the system is an expert in the domain, has full knowledge of the current situation, and is the sole arbiter of correct performance.</Paragraph>
    <Paragraph position="7"> II. Integrate the text plan that implicitly conflicts with the system recommendation as a concession that the user may choose not to accept the recommendation. This strategy is appropriate if the system is an expert in the domain, but the user has better knowledge of the current situation and/or retains responsibility for selecting the best plan of action. Decision support is such an environment. The top half of Figure 6 presents two TraumaTIQ critiques that exhibit implicit conflict, while the bottom part presents the English realization of the integrated text plan, which uses a CONCES-SION relation to achieve coherence.</Paragraph>
    <Paragraph position="8"> III. Present the system recommendation as an alternative to the user plan. This may be appropriate if the parameters indicate the user has more complete knowledge and more expertise.</Paragraph>
    <Paragraph position="9">  Occasionally two text plans may exhibit no conflict, yet the relationships between their communicative goals can be exploited to produce more coherent text. For example, consider the following two individual critiques produced by TraumaTIQ: Caution: do a peritoneal lavage immediately as part of ruling out abdominal bleeding.</Paragraph>
    <Paragraph position="10"> Do not reassess the patient in 6 to 24 hours until after doing a peritoneal lavage. The outcome of the latter may affect the need to do the former.</Paragraph>
    <Paragraph position="11">  While the two critiques do not conflict, RTPI's rules exploit the relation between the communicative goals in their respective text plans to produce a more concise and coherent message. In particular, one of RTPI's rules recognizes the interaction between an initial plan to get the user to perform an action As, and a second plan that gets the user to recognize a dependency between As and another action. This rule creates a text plan for the message: Do a peritoneal lavage immediately as part of ruling out abdominal bleeding. Use the results of the peritoneal lavage to decide whether to reassess the patient in 6 to P4 hours.</Paragraph>
    <Paragraph position="12"> TraumaTIQ critiques: Performing local visual exploration of all abdominal wounds is preferred over doing a peritoneal lavage for ruling out a suspicious abdominal wall injury.</Paragraph>
    <Paragraph position="13"> Please remember to check for laparotomy scars before you do a peritoneal lavage.</Paragraph>
    <Paragraph position="14"> Message from RTPI integrated plan: Performing local visual exploration of all abdominal wounds is preferred over doing a peritoneal lavage for ruling out a suspicious abdominal wall injury. However, if you do a peritoneal lavage, then remember to first check for laparotomy scars.</Paragraph>
    <Section position="1" start_page="516" end_page="516" type="sub_section">
      <SectionTitle>
5.2 Trailing Comments
</SectionTitle>
      <Paragraph position="0"> Occasionally when several text plans are integrated into a single text plan, another text plan that overlaps with the integrated plan will remain outside the new plan because the scoring function for the applicable rule was too low to allow it to combine. This is typically because an effort to integrate such a text plan would create a message so complex that the heuristic deemed it inappropriate.</Paragraph>
      <Paragraph position="1"> However, once concepts have been introduced in the integrated text plan, focusing heuristics (McKeown, 1985) suggest that other text plans containing these concepts be included in the integrated plan as well. Rather than restructure the result of our transformation (against the advice of our heuristic), we append them to the end of the message. Thus we refer to them as trailing comments.</Paragraph>
      <Paragraph position="2"> Unfortunately, when the communicative goal is to get the user to perform an action, trailing comments that refer to such actions have the potential to erroneously suggest new instances of actions. Our solution to this problem is implemented in the text realization templates, where we (1) make the focused action the subject of the sentence, reflecting its given status in the discourse, (2)utilize clue words to call attention to its occurrence earlier in the message and to the new information being conveyed, and (3) subordinate other concepts presented with the focused concept by placing them in a phrase introduced by the cue words &amp;quot;along with&amp;quot;. In one such example from the trauma domain, the main text plan contains the communicative goal of getting the user to perform several actions, including a laparotomy. A SEQUENCE relation is used to adjoin an overlapping text plan as a trailing comment, and this additional communicative goal is realized in English as (clue  Moreover., doing the laparotomy is also indicated, along with repairing the left diaphragm, to treat the lacerated left diaphragm.</Paragraph>
    </Section>
  </Section>
  <Section position="8" start_page="516" end_page="517" type="metho">
    <SectionTitle>
6 Algorithm
</SectionTitle>
    <Paragraph position="0"> RTPI performs rule-based integration of a set of RST-style trees. Rules are applied in an order designed to maximize derived benefit. The system first applies the rules that resolve conflict, since we hypothesize that the presence of conflict will most seriously hamper assimilation of a message. Next, the rules that exploit relations between text plans are tried because they enhance coherence by explicitly connecting different communicative goals. Then the aggregation rules are applied to improve conciseness.</Paragraph>
    <Paragraph position="1"> Finally, the rules for trailing comments reduce the number of disconnected message units.</Paragraph>
    <Paragraph position="2"> The algorithm is both greedy and anytime (Garvey and Lesser, 1994); it takes the best result from a single application of a rule to a set of text plans, and then attempts to further apply rules to the modified set. The rule instantiation with the highest heuristic score is chosen and the rule's operator is applied to the trees using those bindings. Since the rules are designed to apply incrementally to a set, every application of a rule results in an improvement in the conciseness or coherence of the tree set, and the tree set is always a viable set of text plans. The user can thus set a time limit for processing of a tree set, and the algorithm can return an improved set at any time. In practice, however, the processing has never taken more than 1-2 seconds, even for large (25 plans) input sets.</Paragraph>
  </Section>
class="xml-element"></Paper>