Cumulative learning (CL) exploits knowledge acquired on prior tasks to improve learning performance on subsequent related tasks. Consider, for example, a CL system that is learning to play chess. Here, one might expect the system to learn from prior games concepts (e.g., favorable board positions, standard openings, end games, etc.) that can be used for future learning. This is in contrast to base learning (Vilalta & Drissi, 2002) in which a fixed learning algorithm is applied to a single task and performance tends to improve only with more exemplars. So, in CL there tends to be explicit reuse of learned knowledge to constrain new learning, whereas base learning depends entirely upon new external inputs.
Relevant techniques for CL operate over multiple tasks, often at higher levels of abstraction, such as new problem space representations, task-based selection of learning algorithms, dynamic...
- Banerji, R. B. (1964). A Language for the Description of Concepts. General Systems, 9, 135–141.Google Scholar
- Baxter, J. (1995). Learning internal representations. In (COLT): Proceeding of the workshop on computational learning theory, Santa Cruz, California. Morgan Kaufmann.Google Scholar
- Brazdil P., Giraud-Carrier, C., Soares, C., & Vilalta, R. (2009). Metalearning Applications to Data Mining, Springer.Google Scholar
- Caruana, R. (1993). Multitask learning: A knowledge-based source of inductive bias. In Proceedings of the tenth international conference on machine learning, University of Massachusetts, Amherst (pp. 41–48).Google Scholar
- Caruana, R. (1996). Algorithms and applications for multitask learning. In Machine learning: Proceedings of the 13th international conference on machine learning (ICML 1996), Bari, Italy (pp. 87–95). Morgan Kauffmann.Google Scholar
- Cohen, B. L. (1978). A Theory of Structural Concept Formation and Pattern Recognition. Ph.D. Thesis, Department of Computer Science, The University of New South Wales.Google Scholar
- Cohen, B. L., & Sammut, C. A. (1982). Object Recognition and Concept Learning with CONFUCIUS. Pattern Recognition Journal, 15(4), 309–316.Google Scholar
- Mitchell, T. (1980). The need for biases in learning generalizations. Rutgers TR CBM-TR-117.Google Scholar
- Mitchell, T. M., & Thrun, S. B. (1993). Explanation-based neural network learning for robot control. In Hanson, Cowan, & Giles (Eds.), Advances in neural information processing systems 5 (pp. 287–294). San Francisco, CA: Morgan-Kaufmann.Google Scholar
- Nilsson, N. J. (1996). Introduction to machine learning: An early draft of a proposed textbook (p. 12). Online at http://ai.stanford.edu/~nilsson/MLBOOK.pdf. Accessed on July 22, 2010.
- Oblinger, D. (2006). Bootstrapped learning proposer information pamphlet for broad agency announcement 07-04. Online at http://fs1.fbo.gov/EPSData/ODA/Synopses/4965/BAA07-04/BLPIPfinal.pdf.
- Pratt, L. Y., Mostow, J., & Kamm, C. A. (1991). Direct transfer of learned information among neural networks. In Proceedings of the ninth national conference on artificial intelligence (AAAI-91), Anaheim, CA (pp. 584–589).Google Scholar
- Ring, M. (1991). Incremental development of complex behaviors through automatic construction of sensory-motor hierarchies. In Proceedings of the eighth international workshop (ML91), San Mateo, California.Google Scholar
- Silver, D., & Mercer, R. (2002). The task rehearsal method of life-long learning: Overcoming impoverished data. In R. Cohen & B. Spencer (Eds.), Advances in artificial intelligence, 15th conference of the Canadian society for computational studies of intelligence (AI 2002), Calgary, Canada, May 27–29, 2002. Lecture notes in computer science (Vol. 2338, pp. 90–101). London: Springer.Google Scholar
- Silver, D., & Poirier, R. (2005). Requirements for machine lifelong learning. JSOCS Technical Report TR-2005-009, Acadia University.Google Scholar
- Swarup, S., Lakkaraju, K., Ray, S. R., & Gasser, L. (2006). Symbol grounding through cumulative learning. In P. Vogt et al. (Eds.), Symbol grounding and beyond: Proceedings of the third international workshop on the emergence and evolution of linguistic communication, Rome, Italy (pp. 180–191). Berlin: Springer.Google Scholar
- Swarup, S., Mahmud, M. M. H., Lakkaraju, K., & Ray, S. R. (2005). Cumulative learning: Towards designing cognitive architectures for artificial agents that have a lifetime. Tech. Rep. UIUCDCS-R-2005-2514.Google Scholar
- Thrun, S. (1998). Lifelong learning algorithms. In S. Thrun & L. Y. Pratt (Eds.), Learning to learn. Norwell, MA: Kluwer Academic.Google Scholar
- Thrun, S., & Mitchell, T. (1995). Lifelong robot learning. Robotics and Autonomous Systems, 15, 25–46.Google Scholar
- Vilalta, R., & Drissi, Y. (2002). A perspective view and survey of meta-learning. Artificial Intelligence Review, 18, 77–95.Google Scholar