<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0"><channel><description>Computational Neuroscience PhD student @ EPFL 🇨🇭</description><link>https://bsky.app/profile/zihan-wu.bsky.social</link><title>@zihan-wu.bsky.social - Zihan Wu</title><item><link>https://bsky.app/profile/zihan-wu.bsky.social/post/3mpu6jgemr22w</link><description>Another paper on bio-plausible learning rules! &#xA;Together with Ariane and the team, we confirm that local self-supervised learning can learn hidden hierarchical structures of synthetic datasets. On the same datasets, we also clarify the limitation of Direct Feedback Alignment.&#xA;&#xA;[contains quote post or other embedded content]</description><pubDate>04 Jul 2026 23:08 +0000</pubDate><guid isPermaLink="false">at://did:plc:st6lrghulapbtkrybuntuqu7/app.bsky.feed.post/3mpu6jgemr22w</guid></item><item><link>https://bsky.app/profile/zihan-wu.bsky.social/post/3mofw7t6by22a</link><description>Can we match self-supervised backpropagation using local learning rules? We show it is possible in our new paper accepted by ICML. We achieve:&#xA;1. theoretical equivalence to BP in a controlled setup&#xA;2. new SOTA for local learning across image datasets&#xA;3. same performance as BP on multiple datasets</description><pubDate>16 Jun 2026 13:37 +0000</pubDate><guid isPermaLink="false">at://did:plc:st6lrghulapbtkrybuntuqu7/app.bsky.feed.post/3mofw7t6by22a</guid></item></channel></rss>