SueYeon Chung
@sueyeonchung.bsky.social
comp neuro, neural manifolds, neuroAI, physics of learning
assistant professor @ harvard (physics, center for brain science, kempner institute)
proj leader @ Flatiron Institute
https://sites.google.com/site/sueyeonchung/
assistant professor @ harvard (physics, center for brain science, kempner institute)
proj leader @ Flatiron Institute
https://sites.google.com/site/sueyeonchung/
Wow that's really asymmetric! I wonder what causes it 🤔
November 5, 2025 at 9:35 PM
Wow that's really asymmetric! I wonder what causes it 🤔
*Interpretable theory of neural-AI alignment:
- Neural prediction: proceedings.neurips.cc/paper_files/...
- Representation similarity:
arxiv.org/pdf/2502.19648
3/3
- Neural prediction: proceedings.neurips.cc/paper_files/...
- Representation similarity:
arxiv.org/pdf/2502.19648
3/3
June 10, 2025 at 12:15 AM
*Interpretable theory of neural-AI alignment:
- Neural prediction: proceedings.neurips.cc/paper_files/...
- Representation similarity:
arxiv.org/pdf/2502.19648
3/3
- Neural prediction: proceedings.neurips.cc/paper_files/...
- Representation similarity:
arxiv.org/pdf/2502.19648
3/3
*Neuro-inspired & efficient AI:
- MMCR (Maximum Manifold Capacity Representations): proceedings.neurips.cc/paper_files/...
- Contrast-equivariant SSL improves neural-model alignment
proceedings.neurips.cc/paper_files/...
2/3
- MMCR (Maximum Manifold Capacity Representations): proceedings.neurips.cc/paper_files/...
- Contrast-equivariant SSL improves neural-model alignment
proceedings.neurips.cc/paper_files/...
2/3
June 10, 2025 at 12:15 AM
*Neuro-inspired & efficient AI:
- MMCR (Maximum Manifold Capacity Representations): proceedings.neurips.cc/paper_files/...
- Contrast-equivariant SSL improves neural-model alignment
proceedings.neurips.cc/paper_files/...
2/3
- MMCR (Maximum Manifold Capacity Representations): proceedings.neurips.cc/paper_files/...
- Contrast-equivariant SSL improves neural-model alignment
proceedings.neurips.cc/paper_files/...
2/3
Congratulations Mark!
May 23, 2025 at 12:25 PM
Congratulations Mark!
Reposted by SueYeon Chung
TL;DR: Encouraging transformation equivariance in an otherwise standard self-supervised learning setup systematically improves your ability to predict neural responses in area IT.
Paper: openreview.net/pdf?id=AiMs8...
Poster: Poster Session 3 #2204, East Exhibit Hall
Paper: openreview.net/pdf?id=AiMs8...
Poster: Poster Session 3 #2204, East Exhibit Hall
openreview.net
December 12, 2024 at 4:13 PM
TL;DR: Encouraging transformation equivariance in an otherwise standard self-supervised learning setup systematically improves your ability to predict neural responses in area IT.
Paper: openreview.net/pdf?id=AiMs8...
Poster: Poster Session 3 #2204, East Exhibit Hall
Paper: openreview.net/pdf?id=AiMs8...
Poster: Poster Session 3 #2204, East Exhibit Hall