Technology
Maximum-Entropy Fine Grained Classification
Abhimanyu Dubey, Otkrist Gupta, Ramesh Raskar, Nikhil Naik
Neural Information Processing Systems http://nips.cc/
Disentangling Influence: Usingdisentangled representationstoauditmodelpredictions
Motivated by the need to audit complex and black box models, there has been extensiveresearch onquantifying howdatafeatures influence model predictions.
Streaming Bayesian Inference for Crowdsourced Classification
Edoardo Manino, Long Tran-Thanh, Nicholas Jennings
Neural Information Processing Systems http://nips.cc/
Near-OptimalRegretforAdversarialMDPwith DelayedBanditFeedback
The standard assumption in reinforcement learning (RL) is that agents observe feedback for their actions immediately. However, in practice feedback is often observedindelay.
Near-OptimalRegretforAdversarialMDPwith DelayedBanditFeedback
The standard assumption in reinforcement learning (RL) is that agents observe feedback for their actions immediately. However, in practice feedback is often observedindelay.
Harnessing small projectors and multiple views for efficient vision pretraining
We show how this idealized loss can be reformulated to a functionally equivalent loss that is more efficient to compute.
Collaborative Decision Making Using Action Suggestions
Inotherp(ost | st) 1(ost = (st)) where 1 indicator introduce 2 (0,1]. Message Reception Rate Reward Normal Perfect Naive - 1.0 Scaled - 0.99 Noisy - 5.0 Chanceof Random Suggestions Reward Normal Perfect Random Naive - 1.0 Naive - 0.25 Scaled - 0.99 Scaled - 0.25 Noisy - 5.0 Noisy - 1.0 Chanceof R...