## Abstract

Multitask Learning is an approach to inductive transfer that improves generalization by using the domain information contained in the training signals of related tasks as an inductive bias. It does this by learning tasks in parallel while using a shared representation; what is learned for each task can help other tasks be learned better. This paper reviews prior work on MTL, presents new evidence that MTL in backprop nets discovers task relatedness without the need of supervisory signals, and presents new results for MTL with k-nearest neighbor and kernel regression. In this paper we demonstrate multitask learning in three domains. We explain how multitask learning works, and show that there are many opportunities for multitask learning in real domains. We present an algorithm and results for multitask learning with case-based methods like k-nearest neighbor and kernel regression, and sketch an algorithm for multitask learning in decision trees. Because multitask learning works, can be applied to many different kinds of domains, and can be used with different learning algorithms, we conjecture there will be many opportunities for its use on real-world problems.

### References

- Abu-Mostafa, Y. S. (1990). “Learning from Hints in Neural Networks,”
*Journal of Complexity*, 6(2), pp. 192–198.CrossRefGoogle Scholar - Abu-Mostafa, Y. S. (1993). “Hints and the VC Dimension,”
*Neural Computation*, 5(2).Google Scholar - Abu-Mostafa, Y. S. (1995). “Hints,”
*Neural Computation*, 7, pp. 639-671.Google Scholar - Baluja, S. & Pomerleau, D. A. (1995). “Using the Representation in a Neural Network's Hidden Layer for TaskSpecific Focus of Attention,”
*Proceedings of the International Joint Conference on Artificial Intelligence 1995*, IJCAI-95, Montreal, Canada, pp. 133-139.Google Scholar - Baxter, J. (1994). “Learning Internal Representations,” Ph.D. Thesis, The Flinders Univeristy of South Australia.Google Scholar
- Baxter, J. (1995). “Learning Internal Representations,”
*Proceedings of the 8th ACM Conference on Computational Learning Theory*, (COLT-95), Santa Cruz, CA.Google Scholar - Baxter, J. (1996). “A Bayesian/Information Theoretic Model of Bias Learning,”
*Proceedings of the 9th International Conference on Computational Learning Theory*, (COLT-96), Desenzano del Gardo, Italy.Google Scholar - Breiman, L. & Friedman, J. H. (1995). “Predicting Multivariate Responses in Multiple Linear Regression,” ftp://ftp.stat.berkeley.edu/pub/users/breiman/curds-whey-all.ps.Z.Google Scholar
- Caruana, R. (1993). “Multitask Learning: A Knowledge-Based Source of Inductive Bias,”
*Proceedings of the 10th International Conference on Machine Learning*, ML-93, University of Massachusetts, Amherst, pp. 41-48.Google Scholar - Caruana, R. (1994).”Multitask Connectionist Learning,”
*Proceedings of the 1993 Connectionist Models Summer School*, pp. 372-379.Google Scholar - Caruana, R. (1995). “Learning Many Related Tasks at the Same Time with Backpropagation,”
*Advances in Neural Information Processing Systems 7*(Proceedings of NIPS-94), pp. 656-664.Google Scholar - Caruana, R., Baluja, S., & Mitchell, T. (1996). “Using the Future to “Sort Out” the Present: Rankprop and Multitask Learning for Medical Risk Prediction,”
*Advances in Neural Information Processing Systems 8*(Proceedings of NIPS-95), pp. 959-965.Google Scholar - Caruana, R. & de Sa, V. R. (1997). “Promoting Poor Features to Supervisors: Some Inputs Work Better as Outputs,” to appear in
*Advances in Neural Information Processing Systems 9*(Proceedings of NIPS-96).Google Scholar - Caruana, R. (1997). “Multitask Learning,” Ph.D. Thesis, School of Computer Science, Carnegie Mellon University.Google Scholar
- Cooper, G. F. & Herskovits, E. (1992). “A Bayesian Method for the Induction of Probabilistic Networks from Data,”
*Machine Learning*, 9, pp. 309-347.CrossRefGoogle Scholar - Cooper, G. F., Aliferis, C. F., Ambrosino, R., Aronis, J., Buchanan, B. G., Caruana, R., Fine, M. J., Glymour, C., Gordon, G., Hanusa, B. H., Janosky, J. E., Meek, C., Mitchell, T., Richardson, T., and Spirtes, P. (1997). ”An Evaluation of Machine Learning Methods for Predicting Pneumonia Mortality,”
*Artificial Intelligence in Medicine 9*, pp. 107-138.Google Scholar - Craven, M. & Shavlik, J. (1994). “Using Sampling and Queries to Extract Rules from Trained Neural Networks,”
*Proceedings of the 11th International Conference on Machine Learning*, ML-94, Rutgers University, New Jersey, pp. 37-45.Google Scholar - Davis, I. & Stentz, A. (1995). “Sensor Fusion for Autonomous Outdoor Navigation Using Neural Networks,”
*Proceedings of IEEE's Intelligent Robots and Systems Conference*.Google Scholar - Dent, L., Boticario, J., McDermott, J., Mitchell, T., & Zabowski, D. (1992). “A Personal Learning Apprentice,”
*Proceedings of 1992 National Conference on Artificial Intelligence*.Google Scholar - de Sa, V. R. (1994). “Learning Classification with Unlabelled Data,”
*Advances in Neural Information Processing Systems 6*, (Proceedings of NIPS-93), pp. 112-119.Google Scholar - Dietterich, T. G., Hild, H., & Bakiri, G. (1990). “A Comparative Study of ID3 and Backpropagation for English Text-to-speech Mapping,”
*Proceedings of the Seventh International Conference on Artificial Intelligence*, pp. 24-31.Google Scholar - Dietterich, T. G., Hild, H., & Bakiri, G. (1995). “A Comparison of ID3 and Backpropagation for English Text-to-speech Mapping,”
*Machine Learning*, 18(1), pp. 51-80.CrossRefGoogle Scholar - Dietterich, T. G. & Bakiri, G. (1995). “Solving Multiclass Learning Problems via Error-Correcting Output Codes,”
*Journal of Artificial Intelligence Research*, 2, pp. 263-286.Google Scholar - Fine, M. J., Singer, D., Hanusa, B. H., Lave, J., & Kapoor, W. (1993). “Validation of a Pneumonia Prognostic Index Using the MedisGroups Comparative Hospital Database,”
*American Journal of Medicine*.Google Scholar - Fisher, D. H. (1987). “Conceptual Clustering, Learning from Examples, and Inference,”
*Proceedings of the 4th International Workshop on Machine Learning*.Google Scholar - Ghahramani, Z. & Jordan, M. I. (1994). “Supervised Learning from Incomplete Data Using an EM Approach,”
*Advances in Neural Information Processing Systems 6*, (Proceedings of NIPS-93,) pp. 120-127.Google Scholar - Ghahramani, Z. & Jordan, M. I. (1997). “Mixture Models for Learning from Incomplete Data,”
*Computational Learning Theory and Natural Learning Systems, Vol. IV*, R. Greiner, T. Petsche and S.J. Hanson (eds.), Cambridge, MA, MIT Press, pp. 67-85.Google Scholar - Ghosn, J. & Bengio, Y. (1997). “Multi-Task Learning for Stock Selection,” to appear in
*Advances in Neural Information Processing Systems 9*, (Proceedings of NIPS-96).Google Scholar - Hinton, G. E. (1986). “Learning Distributed Representations of Concepts,”
*Proceedings of the 8th International Conference of the Cognitive Science Society*, pp. 1-12.Google Scholar - Holmstrom, L. & Koistinen, P. (1992). “Using Additive Noise in Back-propagation Training,”
*IEEE Transactions on Neural Networks*, 3(1), pp. 24-38.Google Scholar - Jordan, M. & Jacobs, R. (1994). “Hierarchical Mixtures of Experts and the EM Algorithm,”
*Neural Computation*, 6, pp. 181-214.Google Scholar - Koller, D. & Sahami, M. (1996). “Toward Optimal Feature Selection,”
*Proceedings of the 13th International Conference on Machine Learning*, ICML-96, Bari, Italy, pp. 284-292.Google Scholar - Le Cun, Y., Boser, B., Denker, J. S., Henderson, D., Howard, R. E., Hubbard, W., & Jackal, L. D. (1989). ”Backpropagation Applied to Handwritten Zip-Code Recognition,”
*Neural Computation*, 1, pp. 541-551.Google Scholar - Little, R. J. A. & Rubin, D. B. (1987).
*Statistical Analysis with Missing Data*, Wiley, New York.Google Scholar - Liu, H. & Setiono, R. (1996). “A Probibilistic Approach to Feature Selection—A Filter Solution,”
*Proceedings of the 13th International Conference on Machine Learning*,ICML-96, Bari, Italy, pp. 319-327.Google Scholar - Martin, J. D. (1994). “Goal-directed Clustering,”
*Proceedings of the 1994 AAAI Spring Symposium on Goal-directed Learning*.Google Scholar - Martin, J. D. & Billman, D. O. (1994). “Acquiring and Combining Overlapping Concepts,”
*Machine Learning*, 16, pp. 1-37.Google Scholar - Mitchell, T. (1980). “The Need for Biases in Learning Generalizations,” Rutgers University:
*CBM-TR-117*.Google Scholar - Mitchell, T., Caruana, R., Freitag, D., McDermott, J., & Zabowski, D. (1994). “Experience with a Learning Personal Assistant,”
*Communications of the ACM: Special Issue on Agents*, 37(7), pp. 80-91.Google Scholar - Munro, P. W. & Parmanto, B. (1997). “Competition Among Networks Improves Committee Performance,” to appear in
*Advances in Neural Information Processing Systems 9*(Proceedings of NIPS-96).Google Scholar - Omohundro, S. M. (1996). “Family Discovery,”
*Advances in Neural Information Processing Systems 8*, (Proceedings of NIPS-95), pp. 402-408.Google Scholar - O'sullivan, J. & Thrun, S. (1996). “Discovering Structure in Multiple Learning Tasks: The TC Algorithm,”
*Proceedings of the 13th International Conference on Machine Learning*, ICML-96, Bari, Italy, pp. 489-497.Google Scholar - Pomerleau, D. A. (1992). “Neural Network Perception for Mobile Robot Guidance,” Carnegie Mellon University:
*CMU-CS-92-115*.Google Scholar - Pratt, L. Y., Mostow, J., & Kamm, C. A. (1991). “Direct Transfer of Learned Information Among Neural Networks,”
*Proceedings of AAAI-91*.Google Scholar - Pratt, L. Y. (1992). “Non-literal Transfer Among Neural Network Learners,” Colorado School of Mines:
*MCS92-04*.Google Scholar - Quinlan, J. R. (1986). “Induction of Decision Trees,”
*Machine Learning*, 1, pp. 81-106.CrossRefGoogle Scholar - Quinlan, J. R. (1992).
*C4.5: Programs for Machine Learning*, Morgan Kaufman Publishers.Google Scholar - Rumelhart, D. E., Hinton, G. E., & Williams, R. J. (1986). “Learning Representations by Back-propagating Errors,”
*Nature*, 323, pp. 533-536.Google Scholar - Sejnowski, T. J. & Rosenberg, C. R. (1986). “NETtalk: A Parallel Network that Learns to Read Aloud,” John Hopkins:
*JHU/EECS-86/01*.Google Scholar - Sharkey, N. E. & Sharkey, A. J. C. (1992). “Adaptive Generalisation and the Transfer of Knowledge,” University of Exeter:
*R257*.Google Scholar - Sill, J. & Abu-Mostafa, Y. (1997). “Monotonicity Hints,” to appear in
*Neural Information Processing Systems 9*(Proceedings of NIPS-96).Google Scholar - Simard, P., Victorri, B., LeCun, Y., & Denker, J. (1992). “Tangent Prop—A Formalism for Specifying Selected Invariances in an Adaptive Neural Network,”
*Advances in Neural Information Processing Systems 4*(Proceedings of NIPS-91), pp. 895-903.Google Scholar - Spirtes, P., Glymour, C., & Scheines, R. (1993).
*Causation, Prediction, and Search*, Springer-Verlag, New York.Google Scholar - Suddarth, S. C. & Kergosien, Y. L. (1990). “Rule-injection Hints as a Means of Improving Network Performance and Learning Time,”
*Proceedings of the 1990 EURASIP Workshop on Neural Networks*, pp. 120-129.Google Scholar - Suddarth, S. C. & Holden, A. D. C. (1991). “Symbolic-neural Systems and the Use of Hints for Developing Complex Systems,”
*International Journal of Man-Machine Studies*, 35(3), pp. 291-311.Google Scholar - Thrun, S. & Mitchell, T. (1994). “Learning One More Thing,” Carnegie Mellon University:
*CS-94-184*.Google Scholar - Thrun, S. (1995). “Lifelong Learning: A Case Study,” Carnegie Mellon University:
*CS-95-208*.Google Scholar - Thrun, S. (1996a). “Is Learning the N-th Thing Any Easier Than Learning the First?,”
*Advances in Neural Information Processing Systems 8*(Proceedings of NIPS-95), pp. 640-646.Google Scholar - Thrun, S. (1996b).
*Explanation-Based Neural Network Learning: A Lifelong Learning Approach*, Kluwer Academic Publisher.Google Scholar - Tresp, V., Ahmad, S., & Neuneier, R. (1994). “Training Neural Networks with Deficient Data,”
*Advances in Neural Information Processing Systems 6*(Proceedings of NIPS-93), pp. 128-135.Google Scholar - Valdes-Perez, R., & Simon, H. (1994). “A Powerful Heuristic for the Discovery of Complex Patterned Behavior,”
*Proceedings of the 11th International Conference on Machine Learning*, ML-94, Rutgers University, NewJersey, pp. 326-334.Google Scholar - Waibel, A., Sawai, H., & Shikano, K. (1989). “Modularity and Scaling in Large Phonemic Neural Networks,”
*IEEE Transactions on Acoustics, Speech and Signal Processing*, 37(12), pp. 1888-1898.Google Scholar