机器之心整理
参与:杜伟
本周既有 NeurIPS 2019 公布的杰出论文奖和经典论文奖论文,也有 MIT 联合 IBM 推出的使图像识别 SOTA 模型性能下降 40 多个点的 ObjectNet 数据集以及 OpenAI 击败 OG 战队的 Dota 2 智能体论文。
Distribution-Independent PAC Learning of Halfspaces with Massart Noise
Dual Averaging Methods for Regularized Stochastic Learning and Online Optimization
Reinforcement Learning Upside Down: Don't Predict Rewards -- Just Map Them to Actions
ObjectNet: A large-scale bias-controlled dataset for pushing the limits of object recognition models
Dynamic Convolution: Attention over Convolution Kernels
Dota 2 with Large Scale Deep Reinforcement Learning
Point-Voxel CNN for Efficient 3D Deep Learning
作者:Ilias Diakonikolas、Themis Gouleakis、Christos Tzamos
论文链接:https://papers.nips.cc/paper/8722-distribution-independent-pac-learning-of-halfspaces-with-massart-noise
作者:Lin Xiao
论文链接:https://www.microsoft.com/en-us/research/wp-content/uploads/2016/02/xiao10JMLR.pdf
作者:Juergen Schmidhuber
论文链接:https://arxiv.org/abs/1912.02875
作者:Andrei Barbu、David Mayo、Julian Alverio、Julian Alverio 等
论文链接:https://objectnet.dev/objectnet-a-large-scale-bias-controlled-dataset-for-pushing-the-limits-of-object-recognition-models.pdf
项目地址:https://objectnet.dev/
作者:Yinpeng Chen、Xiyang Dai、Mengchen Liu、Dongdong Chen、Lu Yuan、Zicheng Liu
论文链接:https://arxiv.org/pdf/1912.03458.pdf
作者:Christopher Berner、Greg Brockman、Brooke Chan、Brooke Chan 等
论文链接:https://cdn.openai.com/dota-2.pdf
作者:Zhijian Liu、Zhijian Liu、Zhijian Liu、Zhijian Liu
论文链接:https://arxiv.org/pdf/1907.03739.pdf