Hi! 👋 I’m Konpat Preechakul or Ta (ต้า).

I’m a machine learning researcher.

I am curious how can we make machines emerge abstraction from pixels.

I’m a PhD student at Berkeley AI Research (advised by Prof. Alexei Efros).

Before that I worked as a research assistant at VISTEC, Thailand (advised by Prof. Supasorn Suwajanakorn), and studied a master degree at Chulalongkorn University, Thailand (advised by Prof. Boonserm Kijsirikul, Ekapol Chuangsuwanich, and Sira Sriswasdi).

<aside> 📋 More info

🎓 Google scholar

🎓 Semantic scholar

💼 LinkedIn

CV.pdf

</aside>

<aside> 🏠 Contact

✉️ konpat (at) berkeley.edu

🏠 Berkeley, CA

</aside>

Blog archives

Updates

Publications

• 2025 — The Serial Scaling Hypothesis (X thread) Yuxi Liu*, Konpat Preechakul*, Kananart Kuwaranancharoen, Yutong Bai Arxiv

• 2025 — Visual Jenga: Discovering Object Dependencies via Counterfactual Inpainting Anand Bhattad, Konpat Preechakul, Alexei A. Efros Arxiv

• 2023 — DNO: Optimizing Diffusion Noise Can Serve As Universal Motion Priors Korrawe Karunratanakul, Konpat Preechakul, Emre Aksan, Thabo Beeler, Supasorn Suwajanakorn, Siyu Tang CVPR 2024 🎉

• 2023 — GMD: Controllable Human Motion Synthesis via Guided Diffusion Models Korrawe Karunratanakul, Konpat Preechakul, Supasorn Suwajanakorn, Siyu Tang ICCV 2023

• 2021 — Diffusion Autoencoders: Toward a Meaningful and Decodable Representation Konpat Preechakul, Nattanat Chatthee, Suttisak Wizadwongsa, Supasorn Suwajanakorn CVPR 2022 (Oral)

• 2021 — Set Prediction in the Latent Spaces Konpat Preechakul, Chawan Piansaddhayanon, Burin Naowarat, Tirasan Khandhawit, Sira Sriswasdi, Ekapol Chuangsuwanich NeurIPS 2021

• 2020 — Improved image classification explainability with high-accuracy heatmaps Konpat Preechakul, Sira Sriswasdi, Boonserm Kijsirikul, Ekapol Chuangsuwanich iScience

Reviewer services

NeurIPS 2025, CVPR 2025 (Outstanding reviewer), ICCV 2025, CVPR 2024, CVPR 2023

Work experience