Jaihoon Kim
@KimJaihoon
Followers
68
Following
39
Media
25
Statuses
76
🧐 Can we define a better initial prior for Sequential Monte Carlo in reward alignment?. That's exactly what Ψ-Sampler 🔱 does. Check out the paper for details:.📌
arxiv.org
We introduce $Ψ$-Sampler, an SMC-based framework incorporating pCNL-based initial particle sampling for effective inference-time reward alignment with a score-based generative model....
We present our paper ."Ψ-Sampler: Initial Particle Sampling for SMC-Based Inference-Time Reward Alignment in Score Models". Check out more details.arXiv: Website:
0
0
6
📈 Can pretrained flow models generate images from complex compositional prompts—including logical relations and quantities—without further fine-tuning?. 🚀 We have released our code for inference-time scaling for flow models:
github.com
Official code for Inference-Time Scaling for Flow Models via Stochastic Generation and Rollover Budget Forcing - KAIST-Visual-AI-Group/Flow-Inference-Time-Scaling
Inference-Time Scaling for Flow Models via.Stochastic Generation and Rollover Budget Forcing
0
5
29
RT @MinhyukSung: I recently presented our work, “Inference-Time Guided Generation with Diffusion and Flow Models,” at HKUST (CVM 2025 keyno….
0
20
0
RT @yuseungleee: ❗️Vision-Language Models (VLMs) struggle with even basic perspective changes!. ✏️ In our new preprint, we aim to extend th….
0
34
0
RT @MinhyukSung: #ICLR2025 Come join our StochSync poster (#103) this morning! We introduce a method that combines the best parts of Score….
stochsync.github.io
Hello world!
0
7
0
RT @KyeongminYeo: 🎉 Join us tomorrow at the #ICLR2025 poster session to learn about our work, "StochSync," extending pretrained diffusion m….
0
8
0
How can VLM reason in arbitrary perspectives? . 🔥 Perspective-Aware Reasoning in Vision-Language Models via Mental Imagery Simulation proposes a framework that enables spatial reasoning of VLM from arbitrary perspectives.
Perspective-Aware Reasoning in Vision-Language Models via Mental Imagery Simulation
0
0
7
RT @_akhaliq: Perspective-Aware Reasoning in Vision-Language Models via Mental Imagery Simulation
0
18
0
🔥 KAIST Visual AI Group is hiring interns for 2025 Summer. ❓Can non-KAIST students apply? Yes!. ❓Can international students who are not enrolled in any Korean institutions apply? Yes!. More info at .🔗
🚀 We’re hiring!.The KAIST Visual AI Group is looking for Summer 2025 undergraduate interns. Interested in:.🌀 Diffusion / Flow / AR models (images, videos, text, more).🧠 VLMs / LLMs / Foundation models.🧊 3D generation & neural rendering. Apply now 👉
0
1
9
RT @myh4832: 🔥 Grounding 3D Orientation in Text-to-Image 🔥.🎯 We present ORIGEN — the first zero-shot method for accurate 3D orientation gro….
0
18
0
RT @MinhyukSung: Introducing ORIGEN: the first orientation-grounding method for image generation with multiple open-vocabulary objects. It’….
0
5
0
🚀 Check out our inference-time scaling with FLUX. GPT-4o struggles to follow user prompts involving compositional logical relations. Our inference-time scaling enables efficient search to generate samples with precise alignment to the input text. 🔗
GPT-4o vs. Our test-time scaling with FLUX (2/2). GPT-4o cannot precisely understand the text (e.g., misinterpreting “occupying chairs” on the left), while our test-time technique generates an image perfectly aligned with the prompt. Check out more 👇.🌐
0
2
9
RT @TheTuringPost: Inference-time scaling can work for flow models. @kaist_ai proposed 3 key ideas to make it possible:. • SDE-based genera….
0
8
0
RT @MinhyukSung: Unconditional Priors Matter!. The key to improving CFG-based "conditional" generation in diffusion models actually lies in….
0
4
0
RT @yuseungleee: 🔎 Unconditional priors matter!. When fine-tuning diffusion models for conditional tasks, the **unconditional** distributio….
0
13
0
📌 Unconditional Priors Matter!. Fine-tuned diffusion models often degrade in unconditional quality —hurting conditional generation. We show that plugging in richer unconditional priors from other models boosts performance. No retraining needed. 🚀. 🔗:
Unconditional Priors Matter! Improving Conditional Generation of Fine-Tuned Diffusion Models without Additional Training Costs. arXiv: Project:
0
0
6
😎 Inference-Time Ghibling. Our inference-time scaling method aligns flow models to complex prompts with logical relations or object quantities—even where the newly released GPT-4o image generator struggles. 🔗 Project: 📝 Paper:
0
3
5
RT @_akhaliq: Inference-Time Scaling for Flow Models via.Stochastic Generation and Rollover Budget Forcing
0
21
0