JNNS Best Paper Award
The Best Paper Award from Japanese Neural Network Society (JNNS) has been awarded to Prof. Keisuke Toyama (ATR Invited Researcher), Prof. Shin Ishii (ATR Department Head), Atsunori Kanemura (ATR Researcher), and Takeaki Shimokawa (ATR Researcher), who contributed to the following three papers.
- K. Okada, K. Toyama, & Y. Kobayashi, “Different pedunculopontine tegmental neurons signal predicted and actual task rewards, J. Neurosci., 29(15): 4858–70, 2009.
- Reinforcement learning optimizes actions by minimizing predicted reward errors (the difference between predicted and actual rewards). This research has proved the existence of the cell in PPTN that encodes the predicted reward errors, marking a leap for uncovering the neural mechanism of reinforcement learning.
- A. Kanemura, S. Maeda, & S. Ishii, “Superresolution with compound Markov random fields via the variational EM algorithm,” Neural Netw., 22(7): 1025–1034, 2009.
- This research dealt with the reconstruction-type superresolution problem and the accompanying registration problem. We used a compound Markov random field for the prior, and proposed a Bayesian estimation method that marginalizes unknown variables, and showed its effectiveness not only in avoiding overfitting, but also in edge-preserving superresolution.
- S. Shinomoto, H. Kim, T. Shimokawa, N. Matsuno, S. Funahashi, K. Shima, I. Fujita, H. Tamura, T. Doi, K. Kawano, N. Inaba, K. Fukushima, S. Kurkin, K. Kurata, M. Taira, K. Tsutsui, H. Komatsu, T. Ogawa, K. Koida, J. Tanji, & K. Toyama, “Relating neuronal firing patterns to functional differentiation of cerebral cortex,” PLOS Comput. Biol., 5(7): e1000433, 2009.
- By analysing spike times series measured in various areas in the brain, we have uncovered that the irregularity in the signal correlates with the functions including the sensation, association, and movement.
Japanese Neural Network Society (JNNS), List of awardees (in Japanese)