I am a PhD student in neurorobotics and robot learning at the University of Freiburg.
Welcome
This website is an experiment of sorts. Over the last year or so, I have been playing around with logseq as a note taking tool. Looking for a way to better share my notes, I have found the Schrodinger plugin by Aryan Sawhney. This is what happened next.
Discrete Filter Debug
This is a test/examle post. How to >>> import torch >>> import matplotlib.pyplot as plt >>> def plot(pr, sm, po, channel): ... for i, d in enumerate([pr[channel], sm[channel], po[channel], (sm*pr)[channel]]): ... plt.subplot(1, 4, i+1) ... plt.imshow(d) ... plt.show() >>> pr = torch.load('/export/hartzj/TakeLidOffSaucepan/demos_gt/encodings/dcm_10_keypoints_encoder-resnet101bs128split9010/new_ego/2022_05_03-14_34_44/cam_w_prior/80.dat') >>> po = torch.load('/export/hartzj/TakeLidOffSaucepan/demos_gt/encodings/dcm_10_keypoints_encoder-resnet101bs128split9010/new_ego/2022_05_03-14_34_44/cam_w_post/80.dat') >>> sm = torch.load('/export/hartzj/TakeLidOffSaucepan/demos_gt/encodings/dcm_10_keypoints_encoder-resnet101bs128split9010/new_ego/2022_05_03-14_34_44/cam_w_sm/80.dat') Results for sigma=2 Channels: 10, 2, 14 While on 2 it resolves the multimodality, on channel 10 it does not manage to do it. On Channel 14 it further manages to sharpen the cluster, but sometimes it also fails to do that. ...
Curriculum vitae
Currently PhD student at the neurorobotics and robot learning labs, University of Freiburg. Specialized in Reinforcement learning, algorithms, cognitive computation. Research interests Sequential decision problems, policy learning, embodied cognition. Education 2020 - 2022 MSc Computer Science, University of Freiburg (Germany) Minor: Neuroscience 2019 - 2020 Erasmus stay at ESIEE Paris (France) 2015 - 2019 BSc Computer Science, University of Freiburg (Germany) Minor: Cognitive Science ...