Dalle Molle Institute for Artificial Intelligence Research

Dalle Molle Institute for Artificial Intelligence Research
Istituto Dalle Molle di Studi sull'Intelligenza Artificiale
Named after Angelo Dalle Molle
Formation 1988
Type public research institute
Purpose artificial intelligence research
Coordinates 46°01′23″N 8°55′01″E / 46.023°N 8.917°E / 46.023; 8.917Coordinates: 46°01′23″N 8°55′01″E / 46.023°N 8.917°E / 46.023; 8.917
Director
 Jürgen Schmidhuber[1]
Affiliations University of Lugano
SUPSI
Website idsia.ch

The Dalle Molle Institute for Artificial Intelligence Research (Italian: Istituto Dalle Molle di Studi sull'Intelligenza Artificiale; IDSIA) was founded in 1988 by Angelo Dalle Molle through the private Fondation Dalle Molle. In 2000 it became a public research institute, affiliated with the University of Lugano and SUPSI in Ticino, Switzerland. In 1997 it has been ranked as one of the world's top ten AI labs, and one of the world's top four labs in the field of biologically inspired AI.[2]

An iCub humanoid robot in IDSIA's robotics lab

One of the main research themes at IDSIA are the Artificial Ants, which are multi-agent methods inspired by the pheromone-based communication of biological ants, pioneered by former IDSIA senior researcher Marco Dorigo and IDSIA's co-director (since 1995) Luca Maria Gambardella.[3] IDSIA's combinations of Artificial Ants and local search algorithms have become a method of choice for numerous optimization tasks involving some sort of graph, such as vehicle routing and internet routing.[4] The burgeoning activity in this field of swarm intelligence has led to numerous commercial applications and specialized conferences dedicated to ant colony optimization algorithms.[5]

Other major research topics in the group of IDSIA's co-director Juergen Schmidhuber (since 1995) include machine learning algorithms for brain-inspired artificial recurrent neural networks, reinforcement learning, evolutionary algorithms and adaptive robotics, complexity theory, in particular the theory of Kolmogorov complexity,[6] theoretically optimal universal decision makers living in environments obeying arbitrary unknown but computable probabilistic laws, and mathematically sound general problem solvers such as Marcus Hutter's asymptotically fastest algorithm for all well-defined problems.[7]

In 2007 a robotics lab with focus on intelligent and learning robots, especially in the fields of swarm and humanoid robotics, was established.[8]

Between 2009 and 2012, artificial neural networks developed at the institute won eight international competitions in pattern recognition and machine learning.[1] More than a billion people can now use IDSIA's algorithms, for example, through Google's speech recognition for smartphones,[9] which is based on a deep learning recurrent network called Long short term memory[10] trained by Connectionist Temporal Classification.[11]

IDSIA is one of four Swiss research organisations founded by the Dalle Molle foundation, of which three are in the field of artificial intelligence.

See also

References

  1. 1 2 Amara D. Angelica (28 November 2012). How bio-inspired deep learning keeps winning competitions: An interview with Dr. Juergen Schmidhuber on the future of neural networks. Kurzweil Accelerating Intelligence newsletter. Accessed March 2016.
  2. ↑ X-Lab Survey, Business Week Magazine, 1997
  3. ↑ Dorigo, Marco, and Luca Maria Gambardella. "Ant colonies for the travelling salesman problem." BioSystems 43.2 (1997): 73-81.
  4. ↑ Gambardella, Luca Maria, et al. "MACS-VRPTW: A Multiple Ant Colony System for Vehicle Routing Problems with Time Windows." New Ideas in Optimization. 1999.
  5. ↑ Dorigo, Marco, et al., eds. Ant Colony Optimization and Swarm Intelligence: 6th International Conference, ANTS 2008, Brussels, Belgium, September 22–24, 2008, Proceedings. Vol. 5217. Springer, 2008.
  6. ↑ Schmidhuber, Jürgen. "Hierarchies of generalized Kolmogorov complexities and nonenumerable universal measures computable in the limit." International Journal of Foundations of Computer Science 13.04 (2002): 587-612.
  7. ↑ Hutter, Marcus. "The fastest and shortest algorithm for all well-defined problems." International Journal of Foundations of Computer Science 13.03 (2002): 431-443.
  8. ↑ Robotics Lab. IDSIA Robotics Lab. Accessed March 2016.
  9. ↑ H. Sak, A. Senior, K. Rao, F. Beaufays, J. Schalkwyk (2015): Google voice search: faster and more accurate. http://googleresearch.blogspot.ch/2015/09/google-voice-search-faster-and-more.html
  10. ↑ Hochreiter, Sepp, and Jürgen Schmidhuber. "Long short-term memory." Neural computation 9.8 (1997): 1735-1780.
  11. ↑ Graves, A., Fernández, S., Gomez, F., & Schmidhuber, J. (2006, June). Connectionist temporal classification: labelling unsegmented sequence data with recurrent neural networks. In Proceedings of the 23rd international conference on Machine learning (pp. 369-376). ACM.

External links

This article is issued from Wikipedia - version of the Friday, March 11, 2016. The text is available under the Creative Commons Attribution/Share Alike but additional terms may apply for the media files.