presentation v0

Research Challenges for Spoken Language Dialog Systems: 

Research Challenges for Spoken Language Dialog Systems Julie Baca, Ph.D. Assistant Research Professor Center for Advanced Vehicular Systems Mississippi State University Computer Science Graduate Seminar March 3, 2004

Overview: 

Overview Define dialog systems Describe research issues Present current work Give conclusions and discuss future work

What is a Dialog System?: 

What is a Dialog System? Current commercial voice products require adherence to “command and control” language, e.g., User: “Plan Route” Such interfaces are not robust to variations from the fixed words and phrases.

What is a Dialog System?: 

What is a Dialog System? Dialog systems seek to provide a natural conversational interaction between the user and the computer system, e.g., User: “Is there a way I can get to Canal Street from here?

Domains for Dialog Systems: 

Domains for Dialog Systems Travel reservation Weather forecasting In-vehicle driver assistance Call routing On-line learning environments

Dialog Systems: Information Flow: 

Dialog Systems: Information Flow Must model two-way flow of information User-to-system System-to-user

Slide7: 

Dialog System

Research Issues: 

Research Issues Many fundamental problems must be solved for these systems to mature. Three general areas include: Automatic Speech Recognition (ASR) Natural Language Processing (NLP) Human-computer Interaction (HCI)

NLP Issue for Dialog Systems: Semantics: 

NLP Issue for Dialog Systems: Semantics Must assess meaning, not just syntactic correctness. Therefore, must handle ungrammatical inputs, e.g., “Is there a ……where is..…a gas station nearby… …?”

NLP Semantics: 

Employ semantic grammar consisting of case frames with named slots. FRAME: [find] [drive] [find] (*WHERE [arrive_loc]) WHERE (where *[be_verb]) [be_verb] (is)(are)(were) [arriveloc] [*[prep] [placename] *[prep]] [placename] (gas station,hotel,restaurant) [prep] (near, nearest, closest, nearby) NLP Semantics

NLP Issue: Semantic Representation : 

NLP Issue: Semantic Representation Two Approaches: Hand-craft the grammar for the application, using robust parsing to understand meaning [1,2]. Problem: time, expense Use statistical approach, generating initial rules and using annotated tree-banked data to discover the full rule set [3,4]. Problem: annotated training data

NLP Issue: Resolving Meaning Using Context: 

NLP Issue: Resolving Meaning Using Context Must maintain knowledge of the conversational context. After request for nearest gas station, user says, “What is it close to?” Resolving “it” - anaphora Another follow-up by the user, “How about …restaurant?” Resolving “…” with “nearest”- ellipsis

Resolving Meaning: Discourse Analysis: 

Resolving Meaning: Discourse Analysis To resolve such requests, system must track context of the conversation. This is typically handled by a discourse analysis component in the Dialog Manager.

Slide14: 

Dialog System

Dialog Manager: Discourse Analysis: 

Dialog Manager: Discourse Analysis Anaphora resolution approach: Use focus mechanism, assuming conversation has focus [5]. For our example, “gas station” is current focus. But how about: “I’m at Food Max. How do I get to a gas station close to it and a video store close to it?” Problem: Resolving the two “its”.

Dialog Manager: Clarification: 

Dialog Manager: Clarification Often cannot satisfy request in one iteration. The previous example may require clarification from the user, “Do you want to go to the gas station first?”

HCI Issue:System vs. User Initiative: 

HCI Issue: System vs. User Initiative What level of control do you provide user in the conversation?

Mixed Initiative: 

Mixed Initiative Total system initiative provides low usability. Total user initiative introduces higher error rate. Thus, mixed initiative approach, balancing usability and error rate, is taken most often. Allowing user to adapt the level explicitly has also shown merit [6].

HCI Issue: Evaluating Dialog Systems: 

HCI Issue: Evaluating Dialog Systems How to compare and evaluate dialog systems? PARADISE (Paradigm for Dialog Systems Evaluation) has provided a standard framework [7].

PARADISE: Evaluating Dialog Systems: 

PARADISE: Evaluating Dialog Systems Task success Was the necessary information exchanged? Efficiency/Cost Number dialog turns, task completion time Qualitative ASR rejections, timeouts, helps Usability User satisfaction with ASR, task ease, interaction pace, system response

Current Work: 

Current Work Sponsored by CAVS Examining: In-vehicle environment Manufacturing environment Online learning environment Multidisciplinary Team: CS (Baca), ECE (Picone) ECE graduate students Hualin Gao, Theban Stanley CPE UG Patrick McNally

Current Work: In-vehicle Dialog System: 

Current Work: In-vehicle Dialog System Approach Developed prototype in-vehicle system. Allows querying for information in Starkville/MSU area.

Slide23: 

Example frames and associated queries: Drive_Direction: “How can I get from Lee Boulevard to Kroger?” Drive_Address: “Where is the campus bakery?” Drive_Distance: “How far is China Garden?” Drive_Quality: “Find me the most scenic route to Scott Field.” Drive_Turn: “I am on Nash Street. What’s my next turn?” System Architecture DIALOG MANAGER

Slide24: 

Geographic Information System (GIS) contains map routing data for MSU and surrounding area. Dialog manager (DM) first determines the nature of query, then: obtains route data from the GIS database handles presentation of the data to the user Application Development GIS Backend

Slide25: 

Obtained domain-specific data by: Initial data gathering and system testing Retesting after enhancing LM and semantic grammar Initial efforts focused on reducing OOV utterances and parsing errors for NLU module. Application Development Pilot System

Slide26: 

In-Vehicle Dialog System Established a preliminary dialog system for future data collection and research Demonstrated significant domain-specific improvements for in-vehicle dialog systems. Created a testbed for future studies of workforce training applications.

Workforce Training: 

Workforce Training Significant issues in manufacturing environment: Recognition issues: Real-time performance Noisy environments Understanding issues: Multimodal interface for reducing error rate, e.g., voice and tactile. HCI/Human Factors Issues: Response generation to integrate speech and visual output

Online Learning: 

Online Learning Significant issues in online learning environment: Understanding issues: Understanding learner preferences and habits. HCI/Human Factors Issues: Response generation to accommodate learning style. Evaluation.

Research Significance: 

Research Significance Advance the development of dialog systems technology through addressing fundamental issues as they arise in various domains. Potential areas: ASR, NLP, HCI

References: 

References [1] S.J. Young and C.E. Proctor, “The design and implementation of dialogue control in voice operated database inquiry systems,” Computer Speech and Language, Vol.3, no. 4, pp. 329-353, 1992. [2] W. Ward, “Understanding spontaneous speech,” in Proceedings of International Conference on Acoustics, Speech and Signal Processing, Toronto, Canada, 1991, pp. 365-368. [3] R. Pieraccini and E. Levin, “Stochastic representation of semantic structure for speech understanding,” Speech Communication, vol. 11., no.2, pp. 283-288, 1992. [4] Y. Wang and A. Acero, “Evaluation of spoken grammar learning in the ATIS domain,” in Proceedings International Conference on Acoustics, Speech, and Signal Processing, Orlando, Florida, 2002. [5] C. Sidner, “Focusing in the comprehension of definite anaphora,” in Computational Model of Discourse, M. Brady, Berwick, R., eds, 1983, Cambridge, MA, pp. 267-330, The MIT Press. [6] D. Littman and S. Pan, “Empirically evaluating an adaptable spoken language dialog system,” in The Proceedings of International Conference on User Modeling, UM ’99, Banff, Canada, 1999.

References : 

References [7] M. Walker, et al., “PARADISE: A Framework for Evaluating Spoken Dialogue Agents, “ Proceedings of the 35th Annual Meeting of the Association for Computational Linguistics (ACL-97), pp. 271-289, 1997.