Looking for endorsor for arXiv Submission (cs.LG)
Hugging Face Forums [Unofficial]
February 28, 2026
Hi, i am facing issues to find someone to endorse me on arXiv, hope you can help me. I have developed a hybrid dual-system architecture for the Bongard-LOGO benchmark. My approach utilizes a learned adaptive arbitration mechanism that dynamically balances a neural backbone (System 1) with Bayesian program synthesis (System 2). This allows the model to handle the abstract logical “rules” of the benchmark without losing the perceptual flexibility of deep learning.
Key Results:
100.0% accuracy on the Free-form problems (surpassing the 92.1% human expert benchmark).
85.0% average accuracy across the benchmark—a +31.8% improvement over the Meta-Baseline-PS results.
I’ve made the code and weights available on GitHub to ensure the results are fully reproducible.
Project Links:
GitHub: Ayushman125/Bongard-Solver this is an initial draft of paper i want to publish-
Paper Draft: https://www.overleaf.com/read/sdytwvbjbxny#93ea49
arXiv Endorsement Details (cs.LG):
I am preparing to submit this work to arXiv and would be honored if you would consider endorsing the submission.
Link: https://arxiv.org/auth/endorse?x=Y8M9WY
Code: Y8M9WY
I’d be very interested to hear from you.
Thank you for your time. AyushmanSaini
Discussion in the ATmosphere