Four short links: 22 February 2018
Fast Style Transfer, Categorizing Social Media Messages, Finding Secrets, and Misleading Images
- A Closed-form Solution to Photorealistic Image Stylization — Experimental results show that the stylized photos generated by our algorithm are twice more preferred by human subjects in average. Moreover, our method runs 60 times faster than the state-of-the-art approach. Code available. (via Ming-Yu Liu)
- Characterizing Social Media Messages by How They Propagate — Since content information is sparse and noisy on social media, adopting TraceMiner allows you to provide a high degree of classication accuracy even in the absence of content information. Experimental results on real-world data sets show the superiority over state-of-the-art approaches on the task of fake news detection and news categorization. (via Paper a Day)
- GitLeaks — searches full repo history for secrets and keys.
- Obfuscated Gradients — In our recent paper, we evaluate the robustness of eight papers accepted to ICLR 2018 as non-certified white-box-secure defenses to adversarial examples. We find that seven of the eight defenses provide a limited increase in robustness and can be broken by improved attack techniques we develop. It’s very easy to make an image that looks to a human like one thing, but which a deep learning classifier will identify as something else. (via Dan Kaminsky)