Summer Tree is Cyan

Thinking will not overcome fear but action will.

ZstGAN | An Adversarial Approach forUnsupervised Zero-Shot Image-to-Image Translation

Abstract In this work In this workwe, we propose a framework calledZstGAN: By introducing an adversarial training scheme,ZstGAN learns to model each domain with domain-specificfeature distribution ...

Weakly Supervised Classification | Towards Accurate Machine Learning with Low Labeling Costs

Masashi Sugiyama | 弱监督机器学习研究进展

reference link : HPCL-智能计算 · NUDT Title: Weakly Supervised Classification: Towards Accurate Machine Learning with Low Labeling Costs 报告人:Prof. Masashi Sugiyama, The University of Tokyo 报告摘...

Decoupled Novel Object Captioner

(image-text) Abstract In this paper, we introduce the zero-shot novel object caption-ing task where the machine generates descriptions without extratraining sentences about the novel object. To ta...

Generative Adversarial Text to Image Synthesis

Abstract In this work, we develop a novel deeparchitecture and GAN formulation to effectivelybridge these advances in text and image model-ing, translating visual concepts from charactersto pixels....

Dynamic Label Graph Matching for Unsupervised Video Re-Identification

Abstract This pa-per focuses on cross-camera label estimation, which can be subsequently used in feature learning to learn robust re-IDmodels. Specifically, we propose to construct a graph forsamp...

Stepwise Metric Promotion for Unsupervised Video Person Re-identification

Abstract two assumptions two assumptions 1) different video track-lets typically contain different persons。 2)within each tracklet, the frames are mostly of the same per-son. \ Our method is built...

Image-Image Domain Adaptation with Preserved Self-Similarity and Domain-Dissimilarity for Person Re-identification

main methords In our at-tempt, we present a “learning via translation” framework.In the baseline, we translate the labeled images from sourceto target domain in an unsupervised manner. we propose ...

Re-ID 2019 Review

Academic Report

Train/Test on the same domain Part-level features are effective. pose and 3D information are beneficial **Smothness in feature space is beneficial ** Unsupervised domain adaptation Evolutio...

为什么-关于因果关系的新科学 | 01 因果关系之梯

因果推断系列丛书

上帝问的是“什么”,他们回答的却是“为什么”。上帝询问事实,他们回答理由。 而且,两人都深信,列举原因可以以某种方式美化他们的行为。他们是从哪里得到这样的想法的? 人类祖先想象不存在之物的能力是一切的关键,正是这种能力让他们得以交流得更加顺畅。在获得这种能力之前,他们只相信自己的直系亲属或者本部落的人。而此后,信任就因共同的幻想 (例如信仰无形但可想象的神,信仰来世,或者信仰领袖的神性)和...

为什么-关于因果关系的新科学 | 导言

因果推断系列丛书

[美]朱迪亚·珀尔 [美]达纳·麦肯齐 著 推荐序 以平实的话语介绍了因果推断的理论建构 对渴望了解因果推断的人们来说,它既是因果关系科学的入门书,又是关于这门学问从萌发到蓬勃发展的一部简史,其中不乏对当前的人工智能发展现状的反思和对未来人工智能发展方向的探索。 the book of why the book of change 这样一本重量级的科普读物,即便是对于一 位专门从事人...