https://yrxxxxxxxx1007.github.io/Ruixuan YangHugoBlox Kit (https://hugoblox.com)en-usSat, 26 Sep 2026 00:00:00 +0000https://yrxxxxxxxx1007.github.io/media/icon_hu_744755e2168fe2f4.pnghttps://yrxxxxxxxx1007.github.io/https://yrxxxxxxxx1007.github.io/publications/mind/Wed, 07 Jan 2026 00:00:00 +0000https://yrxxxxxxxx1007.github.io/publications/mind/<h2 id="overview">Overview</h2> <p>Aligning diverse reasoning perspectives with the student&rsquo;s evolving capabilities to move beyond passive imitation.</p> <p> <figure > <div class="flex justify-center "> <div class="w-full" > <img alt="Overview of the MIND capability-aware multi-perspective distillation framework" srcset="https://yrxxxxxxxx1007.github.io/publications/mind/featured_hu_f2f32bf9b337dc76.webp 320w, https://yrxxxxxxxx1007.github.io/publications/mind/featured_hu_cfce2cdb414a3285.webp 480w, https://yrxxxxxxxx1007.github.io/publications/mind/featured_hu_8c00ebec3171fcfd.webp 760w" sizes="(max-width: 480px) 100vw, (max-width: 768px) 90vw, (max-width: 1024px) 80vw, 760px" src="https://yrxxxxxxxx1007.github.io/publications/mind/featured_hu_f2f32bf9b337dc76.webp" width="760" height="373" loading="lazy" data-zoomable /></div> </div></figure> </p> <p><strong>ACL 2026 · Main Conference · Long Paper.</strong></p> <p><a href="https://aclanthology.org/2026.acl-long.2020/" target="_blank" rel="noopener">Read the paper</a> · <a href="https://aclanthology.org/2026.acl-long.2020.pdf" target="_blank" rel="noopener">Download PDF</a></p>https://yrxxxxxxxx1007.github.io/publications/compact/Tue, 20 Jan 2026 00:00:00 +0000https://yrxxxxxxxx1007.github.io/publications/compact/<h2 id="overview">Overview</h2> <p>Combining complementary teachers through student-aware supervision for more effective reasoning distillation.</p> <p> <figure > <div class="flex justify-center "> <div class="w-full" > <img alt="Overview of the COMPACT multi-teacher reasoning distillation framework" srcset="https://yrxxxxxxxx1007.github.io/publications/compact/featured_hu_66dcdc79d7a93cf9.webp 320w, https://yrxxxxxxxx1007.github.io/publications/compact/featured_hu_5bf7e61fa0d1cf66.webp 480w, https://yrxxxxxxxx1007.github.io/publications/compact/featured_hu_c3358d2f7d0ac821.webp 760w" sizes="(max-width: 480px) 100vw, (max-width: 768px) 90vw, (max-width: 1024px) 80vw, 760px" src="https://yrxxxxxxxx1007.github.io/publications/compact/featured_hu_66dcdc79d7a93cf9.webp" width="760" height="360" loading="lazy" data-zoomable /></div> </div></figure> </p> <p>Accepted to <strong>IJCAI 2026</strong>. Jin Cui and Jiaqi Guo contributed equally.</p> <p><a href="https://arxiv.org/abs/2601.13992" target="_blank" rel="noopener">Read the paper</a> · <a href="https://arxiv.org/pdf/2601.13992" target="_blank" rel="noopener">Download PDF</a></p>https://yrxxxxxxxx1007.github.io/publications/echo/Sat, 09 May 2026 00:00:00 +0000https://yrxxxxxxxx1007.github.io/publications/echo/<h2 id="overview">Overview</h2> <p>Organizing robot experiences in a continuous hierarchical memory to support retrieval and generalization in long-horizon tasks.</p> <p> <figure > <div class="flex justify-center "> <div class="w-full" > <img alt="ECHO architecture integrating a hierarchical memory system with a vision-language-action model" srcset="https://yrxxxxxxxx1007.github.io/publications/echo/featured_hu_4bc5bd9f0febb986.webp 320w, https://yrxxxxxxxx1007.github.io/publications/echo/featured_hu_5d41ba5c8a0807d0.webp 480w, https://yrxxxxxxxx1007.github.io/publications/echo/featured_hu_194eb872fdff28a3.webp 760w" sizes="(max-width: 480px) 100vw, (max-width: 768px) 90vw, (max-width: 1024px) 80vw, 760px" src="https://yrxxxxxxxx1007.github.io/publications/echo/featured_hu_4bc5bd9f0febb986.webp" width="760" height="372" loading="lazy" data-zoomable /></div> </div></figure> </p> <p><strong>NeurIPS 2026 · Accepted.</strong></p> <p><a href="https://arxiv.org/abs/2605.10993" target="_blank" rel="noopener">Read the paper</a> · <a href="https://arxiv.org/pdf/2605.10993" target="_blank" rel="noopener">Download PDF</a></p>https://yrxxxxxxxx1007.github.io/publications/lt-opd/Sat, 26 Sep 2026 00:00:00 +0000https://yrxxxxxxxx1007.github.io/publications/lt-opd/<h2 id="overview">Overview</h2> <p>Recovering multimodal capabilities under extreme visual-token reduction through on-policy self-distillation and a progressive token-budget curriculum.</p> <p> <figure > <div class="flex justify-center "> <div class="w-full" > <img alt="LT-OPD framework with a low-token student, a frozen full-token teacher, and a token-budget curriculum" srcset="https://yrxxxxxxxx1007.github.io/publications/lt-opd/featured_hu_6aa0ccbbac72b2d9.webp 320w, https://yrxxxxxxxx1007.github.io/publications/lt-opd/featured_hu_243610fc64eb846f.webp 480w, https://yrxxxxxxxx1007.github.io/publications/lt-opd/featured_hu_2991ac9026c4f9fc.webp 760w" sizes="(max-width: 480px) 100vw, (max-width: 768px) 90vw, (max-width: 1024px) 80vw, 760px" src="https://yrxxxxxxxx1007.github.io/publications/lt-opd/featured_hu_6aa0ccbbac72b2d9.webp" width="760" height="382" loading="lazy" data-zoomable /></div> </div></figure> </p> <p><strong>arXiv 2026 · Preprint.</strong></p> <p><a href="https://arxiv.org/abs/2609.32353" target="_blank" rel="noopener">Read the paper</a> · <a href="https://arxiv.org/pdf/2609.32353" target="_blank" rel="noopener">Download PDF</a> · <a href="https://github.com/Yrxxxxxxxx1007/LT-OPD" target="_blank" rel="noopener">Code</a></p>