A network approach to probabilistic information retrieval

ABSTRACT

In this article we show how probabilistic information retrieval based on document components may be implemented as a feedforward (feedbackward) artificial neural network. The network supports adaptation of connection weights as well as the growing of new edges between queries and terms based on user relevance feedback data for training, and it reflects query modification and expansion in information retrieval. A learning rule is applied that can also be viewed as supporting sequential learning using a harmonic sequence learning rate. Experimental results with four standard small collections and a large Wall Street Journal collection (173,219 documents) show that performance of feedback improves substantially over no feedback, and further gains are obtained when queries are expanded with terms from the feedback documents. The effect is much more pronounced in small collections than in the large collection. Query expansion may be considered as a tool for both precision and recall enhancement. In particular, small query expansion levels of about 30 terms can achieve most of the gains at the low-recall high-precision region, while larger expansion levels continue to provide gains at the high-recall low-precision region of a precision recall curve.

REFERENCES

• 1: BEIN, J. AND SMOLENSKY, P. 1988. Application of the interactive activation model to document retrieval. Tech. Rep. CU-CS-405-88, Dept. of Computer Science, Univ. of Colorado, Boulder, Colo.
• 2: Abraham Bookstein, A comparison of two weighting schemes for Boolean retrieval, Proceedings of the 3rd annual ACM conference on Research and development in information retrieval, p.23-34, June 23-27, 1980, Cambridge, England
• 3: BOOKSTEIN, A. AND SWANSON, D. R. 1975. A decision theoretic foundation for indexing. In J. Am. Soc. Inf. Sct. 26, 45-50.
• 4: R. J. Brachman , D. L. McGuinness, Knowledge representation, connectionism and conceptual retrieval, Proceedings of the 11th annual international ACM SIGIR conference on Research and development in information retrieval, p.161-174, May 1988, Grenoble, France [doi>10.1145/62437.62448]
• 5: BRAUEN, R. L. 1971. Document vector modification. In SMART System--Experiments in Automatic Document Processing, G. Salton, Ed. Prentice-Hall, Englewood Cliffs, N.J., 456-484.
• 6: Steven C. Cater , Donald H. Kraft, A generalization and clarification of the Waller-Kraft wish list, Information Processing and Management: an International Journal, v.25 n.1, p.15-25, 1989 [doi>10.1016/0306-4573(89)90088-5]
• 7: William S. Cooper, Some inconsistencies and misnomers in probabilistic information retrieval, Proceedings of the 14th annual international ACM SIGIR conference on Research and development in information retrieval, p.57-61, October 13-16, 1991, Chicago, Illinois, United States [doi>10.1145/122860.122866]
• 8: William S. Cooper , Fredric C. Gey , Daniel P. Dabney, Probabilistic retrieval based on staged logistic regression, Proceedings of the 15th annual international ACM SIGIR conference on Research and development in information retrieval, p.198-210, June 21-24, 1992, Copenhagen, Denmark [doi>10.1145/133160.133199]
• 9: CROFT, W. B. AND HARPER, D. J. 1979. Using probabilistic models of document retrieval without relevance information. J. Doc. 35, 285-295.
• 10: DEERWESTER, S., DUMAIS, S. T., FURNAS, G. W., LANDAUER, T. K., AND HARSHMAN, R. 1990. Indexing by latent semantic analysis. J. Am. Soc. Inf. Sct. 41,391-407.
• 11: Norbert Fuhr, Optimum polynomial retrieval functions based on the probability ranking principle, ACM Transactions on Information Systems (TOIS), v.7 n.3, p.183-204, July 1989 [doi>10.1145/65943.65944]
• 12: GALLANT, S. I. 1991. Context vector representation for document retrieval. In AAAI-91: Natural Language Text Retrteval Workshop (July 15). AAAI, San Francisco, Calif.
• 13: David Haines , W. Bruce Croft, Relevance feedback and inference networks, Proceedings of the 16th annual international ACM SIGIR conference on Research and development in information retrieval, p.2-11, June 27-July 01, 1993, Pittsburgh, Pennsylvania, United States [doi>10.1145/160688.160689]
• 14: Donna Harman, Relevance feedback revisited, Proceedings of the 15th annual international ACM SIGIR conference on Research and development in information retrieval, p.1-10, June 21-24, 1992, Copenhagen, Denmark [doi>10.1145/133160.133167]
• 15: HAP~AN, D., Ed. 1993. The 1st Text Retrieval Conference (TREC-1) (Mar.). Special Pub. 500-207, National Institute of Standards and Technology, Gaithersburg, Md.
• 16: T. Kohonen, Self-organization and associative memory: 3rd edition, Springer-Verlag New York, Inc., New York, NY, 1989
• 17: K. L. Kwok, A neural network for probabilistic information retrieval, Proceedings of the 12th annual international ACM SIGIR conference on Research and development in information retrieval, p.21-30, June 25-28, 1989, Cambridge, Massachusetts, United States
• 18: KWOK, K. L. 1990a. Application of neural network to information retrieval. In Proceedings of the International Joint Conference on Neural Networks, M. Caudill, Ed. Vol. 2. Erlbaum, Hillsdale, N.J., 623-626.
• 19: K. L. Kwok, Experiments with a component theory of probabilistic information retrieval based on single terms as document components, ACM Transactions on Information Systems (TOIS), v.8 n.4, p.363-386, Oct. 1990 [doi>10.1145/102675.102677]
• 20: K. L. Kwok , William Kuan, Experiments with document components for indexing and retrieval, Information Processing and Management: an International Journal, v.24 n.4, p.405-417, May 1988 [doi>10.1016/0306-4573(88)90044-1]
• 21: KWOK, K. L., PAPADOPOLOUS, L., ANn KWAN, Y. Y. 1993. Retrieval experiments with a large collection using PIRCS. In The 1st Text REtrieval Conference (TREC-1). Special Pub. 500-267, National Institute of Standards and Technology, Gaithersburg, Md.
• 22: M. E. Maron , J. L. Kuhns, On Relevance, Probabilistic Indexing and Information Retrieval, Journal of the ACM (JACM), v.7 n.3, p.216-244, July 1960 [doi>10.1145/321033.321035]
• 23: Brian P. McCune , Richard M. Tong , Jeffrey S. Dean , Daniel G. Shapiro, RUBRIC: a system for rule-based information retrieval, IEEE Transactions on Software Engineering, v.11 n.9, p.939-945, Sept. 1985 [doi>10.1109/TSE.1985.232827]
• 24: MOZER, M. C. 1984. Inductive information retrieval using parallel distributed computation. ICS Tech. Rep. 8406, Univ. of California, San Diego, Calif.
• 25: RADECKI, T. 1979. Fuzzy-set theoretical approach to document retrieval. Inf. Proc. Mgmt. 15, 247-259.
• 26: ROBERTSON, S. E. 1977. The probability ranking principle in IR. J. Doc. 33, 294-304.
• 27: ROBERTSON, S. E. AND SPARCK JONES, K. 1976. Relevance weighting of search terms. J. Am. Soc. Inf. Sci. 27, 129-146.
• 28: ROBERTSON, S. E., MARON, M. E., AND COOPER, W. S. 1982. Probability of relevance: A unification of two competing models for document retrieval. Inf. Tech. Res. Devel. i, 1-21.
• 29: ROCCHIO, J. J., JR. 1971. Relevance feedback in information retrieval. In The Smart System--Experiments in Automatic Document Processing, G. Salton, Ed. Prentice-Hall, Englewood Cliffs, N.J., 313-323.
• 30: Daniel E. Rose , Richard K. Belew, A connectionist and symbolic hybrid for improving legal research, International Journal of Man-Machine Studies, v.35 n.1, p.1-33, July 1991 [doi>10.1016/S0020-7373(07)80006-X]
• 31: David E. Rumelhart , James L. McClelland , CORPORATE PDP Research Group, Parallel distributed processing: explorations in the microstructure of cognition, vol. 1: foundations, MIT Press, Cambridge, MA, 1986
• 32: Gerald Salton, Automatic text processing, Addison-Wesley Longman Publishing Co., Inc., Boston, MA, 1988
• 33: SALTON, G. AND BUCKLEY, C. 1990. Improving retrieval performance by relevance feedback. J. Am. Soc. Inf. Sci. 41,288-297.
• 34: Gerard Salton , Edward A. Fox , Harry Wu, Extended Boolean information retrieval, Communications of the ACM, v.26 n.11, p.1022-1036, Nov. 1983 [doi>10.1145/182.358466]
• 35: SMEATON, A. F. AND VAN RIJSBERGEN, C. J. 1983. The retrieval effects of query expansion on a feedback document retrieval system. Comput. J. 26, 239 246.
• 36: Philip J. Smith , Steven J. Shute , Beb Galdes , Mark H. Chignell, Knowledge-based search tactics for an intelligent intermediary system, ACM Transactions on Information Systems (TOIS), v.7 n.3, p.246-270, July 1989 [doi>10.1145/65943.65947]
• 37: SPECK JONES, K. 1972. A statistical interpretation of term specificity and its application in retrieval. J. Doc. 8, 11-21.
• 38: Howard Turtle , W. Bruce Croft, Evaluation of an inference network-based retrieval model, ACM Transactions on Information Systems (TOIS), v.9 n.3, p.187-222, July 1991 [doi>10.1145/125187.125188]
• 39: C. J. Van Rijsbergen, Information Retrieval, Butterworth-Heinemann, Newton, MA, 1979
• 40: C. J. van Rijsbergen, Towards an information logic, Proceedings of the 12th annual international ACM SIGIR conference on Research and development in information retrieval, p.77-86, June 25-28, 1989, Cambridge, Massachusetts, United States
• 41: WONG, S. K. M., YAO, Y. Y., SALTON, G., AND BUCKLEY, C. 1991. Evaluation of an adaptive linear model. J. Am. Soc. Inf. Sci. 42, 723-730.
• 42: S. K.M. Wong , W. Ziarko , V. V. Raghavan , P. C.N. Wong, On modeling of information retrieval concepts in vector spaces, ACM Transactions on Database Systems (TODS), v.12 n.2, p.299-321, June 1987 [doi>10.1145/22952.22957]
• 43: C. T. Yu , G. Salton, Precision Weighting—An Effective Automatic Indexing Method, Journal of the ACM (JACM), v.23 n.1, p.76-88, Jan. 1976 [doi>10.1145/321921.321930]
• 44: Yu, C. T., BUCKLEY, C., L~, K., AND SALTON, G. 1983. A generalized term dependence model in information retrieval. Inf. Tech. Res. Devel. 2, 129-154.

INDEX TERMS

Primary Classification:
  H. Information Systems
  H.3 INFORMATION STORAGE AND RETRIEVAL
      H.3.3 Information Search and Retrieval
          Subjects: Retrieval models

Additional Classification:
  E. Data

  H. Information Systems
  H.3 INFORMATION STORAGE AND RETRIEVAL
      H.3.1 Content Analysis and Indexing
          Subjects: Indexing methods

  I. Computing Methodologies
  I.2 ARTIFICIAL INTELLIGENCE
      I.2.6 Learning
          Subjects: Connectionism and neural nets

General Terms:
Algorithms, Experimentation, Performance

Keywords:
artificial neural networks, document-focused and query-focused relevance feedback, indexing and retrieval, item self-learning, learning, probabilistic indexing, probabilistic retrieval, query expansion, training