Natural language generation for sponsored-search advertisements

ABSTRACT

In sponsored search, advertisers bid on phrases representative of offered products or services. For large advertisers, these phrases often come from quasi-algorithmically generated lists of thousands of terms prone to poor linguistic construction. A bidded term by itself is usually unsuitable for direct insertion into an ad copy template; it must be rephrased and capitalized properly to fit the template, possibly with additional language to avoid semantic ambiguity. We develop a natural language generation system to automate these steps, preparing a list of terms for insertion into an ad template. For each input term, our system first finds a proper word ordering by mining a corpus of Web search query logs. Next it determines whether the term is ambiguous and--if semantics dictate--attaches a clarifying modifier culled from query logs. Finally, it applies proper capitalization by analyzing pages from Web search engine results. Each step yields a plausible set of displayable forms from which a machine-learned model selects the best. The models are trained and tested on a large set of human-labeled data. The overall system significantly outperforms baseline systems that use simple heuristics.

REFERENCES

• 1: H. Akaike. A new look at the statistical model identification. IEEE Transactions on Automatic Control, 19(6):716--723, 1974.
• 2: Iñaki Alegria , Bertol Arrieta , Arantza Diaz de Ilarraza , Eli Izagirre , Montse Maritxalar, Using machine learning techniques to build a comma checker for Basque, Proceedings of the COLING/ACL on Main conference poster sessions, p.1-8, July 17-18, 2006, Sydney, Australia
• 3: M. Arellano. Panel Data Econometrics. Oxford University Press, Oxford, UK, 2003.
• 4: C. Chelba and A. Acero. Adaptation of maximum entropy capitalizer: Little data can help a lot. Computer Speech & Language, 20(4):382--399, 2006.
• 5: D. Conway. An algorithmic approach to english pluralization. Proceedings of the Second Annual Perl Conference, 1998.
• 6: Barbara Di Eugenio , Johanna D. Moore , Massimo Paolucci, Learning features that predict cue usage, Proceedings of the 35th Annual Meeting of the Association for Computational Linguistics and Eighth Conference of the European Chapter of the Association for Computational Linguistics, p.80-87, July 07-12, 1997, Madrid, Spain
• 7: D. C. Fain and J. O. Pedersen. Sponsored search: A brief history. Proceedings of the ACM Conference on Electronic Commerce: Second Workshop on Sponsored Search Auctions, 2006.
• 8: D. Hardt. Comma checking in danish. Proceedings of the Corpus Linguistics 2001 Conference, pages 266--271, 2001.
• 9: Bernard J. Jansen , Amanda Spink , Tefko Saracevic, Real life, real users, and real needs: a study and analysis of user queries on the web, Information Processing and Management: an International Journal, v.36 n.2, p.207-227, Jan.1.2000 [doi>10.1016/S0306-4573(99)00056-4]
• 10: D. McFadden. Conditional logit analysis of qualitative choice behavior. In P. Zarembka, editor, Frontiers in Econometrics. Academic Press, New York, 1974.
• 11: J. Oberlander and C. Brew. Stochastic text generation. Philosophical Transactions: Mathematical, Physical and Engineering Science, 358:1373--1387, 2000.
• 12: M. Porter. An algorithm for suffix stripping. Program, 14(3):130--137, 1980.
• 13: Dragomir R. Radev , Kathleen R. McKeown, Generating natural language summaries from multiple on-line sources, Computational Linguistics, v.24 n.3, September 1998
• 14: C. Reed and D. Long. Generating punctuation in written arguments. Technical Report RN/97/157, Department of Computer Science, University College, London, UK, 1997.
• 15: Christoph Tillmann , Hermann Ney, Word reordering and a dynamic programming beam search algorithm for statistical machine translation, Computational Linguistics, v.29 n.1, p.97-133, March 2003 [doi>10.1162/089120103321337458]
• 16: K. Train. Discrete Choice Methods with Simulation. Cambridge University Press, Cambridge, UK, 2003.

INDEX TERMS

Primary Classification:
  I. Computing Methodologies
  I.7 DOCUMENT AND TEXT PROCESSING
      I.7.m Miscellaneous

General Terms:
Algorithms, Economics, Languages

Keywords:
automation, machine learning, natural language generation, natural language processing, pay-per-click, query log, sponsored search