Laurent Perrinet<ul><li>Quantitative results assessing the role of <a href="https://neuromatch.social/tags/homeostasis" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>homeostasis</span></a> in the <a href="https://neuromatch.social/tags/unsuperviseelearning" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>unsuperviseelearning</span></a> of <a href="https://neuromatch.social/tags/spatiotemporal" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>spatiotemporal</span></a> <a href="https://neuromatch.social/tags/patterns" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>patterns</span></a> and the performance of the <a href="https://neuromatch.social/tags/neuralnetwork" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>neuralnetwork</span></a> on different <a href="https://neuromatch.social/tags/datasets" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>datasets</span></a> : Poker-DVS, N-MNIST and DVS Gesture for different precisions of the temporal and spatial information.</li></ul><p>Here, we display the performance of the algorithm on the DVSgesture dataset. For this gesture recognition task, the online HOTS accuracy remains close to the chance level for about 100 events. More evidence needs to be accumulated, and then the accuracy increases monotonically, outperforming the previous method after about 10.000 events (at an average of 9.3% of the number of events in the sample) :</p>