Author name cluster

Daewon Chae

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

5 papers

2 author rows

AAMAS Conference 2026 Conference Paper

Towards Failure-Resilient Lifelong Learning Agents through Scene Graph-Guided Proactive Replanning

Che Rin Yu
Daewon Chae
Dabin Seo
Sangwon Lee
Hyeongwoo IM
Jinkyu Kim

When humans perform everyday tasks, we naturally adjust our actions based on the current state of the environment. For example, if weplantoheatabowlofsoupinthemicrowaveandseethatthere’s already a plate inside, we first remove it before proceeding. However, many autonomous robots lack this adaptive awareness. They often follow pre-planned actions that may overlook subtle yet critical changes in the scene, which can result in actions being executed under outdated assumptions and eventual failure. While replanning is critical for robust autonomy, most existing methods respond only after failures occur, when recovery may be inefficient or infeasible. In this work, we present a proactive replanning framework that anticipates and prevents failures before action execution. The key idea is to learn visual preconditions from successful demonstrations. Before each subtask, the system builds a scene graph from RGB-D observations and compares it with an expanding buffer of reference graphs from successful trials. When the similarity falls below a threshold, a lightweight reasoning module diagnoses the discrepancy and generates a corrective sub-plan. As the experience buffer grows, the system scales and generalizes more efficiently, highlighting a progressive path toward more autonomous and effective replanning. Experiments in both the AI2-THOR simulator and real-world platforms show that our approach improves task success and execution performance compared to baselines.

PDF

AAAI Conference 2025 Conference Paper

DiffExp: Efficient Exploration in Reward Fine-tuning for Text-to-Image Diffusion Models

Daewon Chae
June Suk Choi
Jinkyu Kim
Kimin Lee

Fine-tuning text-to-image diffusion models to maximize rewards has proven effective for enhancing model performance. However, reward fine-tuning methods often suffer from slow convergence due to online sample generation. Therefore, obtaining diverse samples with strong reward signals is crucial for improving sample efficiency and overall performance. In this work, we introduce DiffExp, a simple yet effective exploration strategy for reward fine-tuning of text-to-image models. Our approach employs two key strategies: (a) dynamically adjusting the scale of classifier-free guidance to enhance sample diversity, and (b) randomly weighting phrases of the text prompt to exploit high-quality reward signals. We demonstrate that these strategies significantly enhance exploration during online sample generation, improving the sample efficiency of recent reward fine-tuning methods, such as DDPO and AlignProp.

PDF Details DOI

TMLR Journal 2025 Journal Article

Encoder-only Next Token Prediction

Ethan Ewer
Daewon Chae
Thomas Zeng
Jinkyu Kim
Kangwook Lee

Next-token prediction is conventionally done using decoder-only Transformers with causal attention, as this approach allows for efficient reuse of keys and values. What if we were not compute-limited, should we still use decoder-only Transformers? In this work, we introduce Encoder-only Next Token Prediction (ENTP). We explore the differences between ENTP and decoder-only Transformers in expressive power and complexity, highlighting potential advantages of ENTP in settings with unbounded compute. We introduce the $\operatorname{Count3}$ task and show, both theoretically and experimentally, that while ENTP can perform this task easily, a decoder-only Transformer cannot. Finally, we empirically demonstrate the superior performance of ENTP across representative tasks where next-token prediction based Transformers can be evaluated, including addition, in-context learning, and language modeling.

PDF Details

TMLR Journal 2025 Journal Article

Hard-Negative Prototype-Based Regularization for Few-Shot Class-Incremental Learning

Seongbeom Park
Hyunju Yun
Daewon Chae
Sungyoon Kim
Suhong Moon
Minwoo Kang
Seunghyun Park
Jinkyu Kim

Few-shot class-incremental learning (FSCIL)---involving abundant base training data followed by novel classes with limited labeled samples---poses challenges such as catastrophic forgetting and overfitting, leading to significant performance degradation across incremental sessions. As a remedy, recent work focuses on minimizing the interference of embeddings between base and incremental classes. However, previous studies have not explicitly considered variation in discriminative difficulty across samples and classes, leaving room for improvement: we observe that hard-negative (i.e., difficult to discriminate from the label) samples and classes significantly affect FSCIL performance, whereas easy ones have little impact. To this end, we propose a hard-negative prototype-based regularization approach that enhances discrimination between similar classes by imposing a penalty margin between each sample and its most similar class prototypes based on cosine similarity. To select hard-negative prototypes, we explore two distinct mining strategies: dynamic selection that leverages the model's decision boundary, and static selection that utilizes a pre-defined class-wise similarity matrix derived from external sources such as pre-trained models. We evaluate our approach on three widely used benchmarks, miniImageNet, CIFAR100, and CUB200, achieving state-of-the-art performance on each. Comprehensive analyses demonstrate that our proposed method enhances intra-class cohesion and inter-class separability of embeddings, both of which are crucial for FSCIL to better accommodate novel classes.

PDF Details

ICML Conference 2025 Conference Paper

VersaPRM: Multi-Domain Process Reward Model via Synthetic Reasoning Data

Thomas Zeng 0003
Shuibai Zhang
Shutong Wu
Christian Classen
Daewon Chae
Ethan Ewer
Minjae Lee
Heeju Kim

Process Reward Models (PRMs) have proven effective at enhancing mathematical reasoning for Large Language Models (LLMs) by leveraging increased inference-time computation. However, they are predominantly trained on mathematical data and their generalizability to non-mathematical domains has not been rigorously studied. In response, this work first shows that current PRMs have poor performance in other domains. To address this limitation, we introduce VersaPRM, a multi-domain PRM trained on synthetic reasoning data generated using our novel data generation and annotation method. VersaPRM achieves consistent performance gains across diverse domains. For instance, in the MMLU-Pro category of Law, VersaPRM via weighted majority voting, achieves a 7. 9% performance gain over the majority voting baseline–surpassing Qwen2. 5-Math-PRM’s gain of 1. 3%. We further contribute to the community by open-sourcing all data, code and models for VersaPRM.

Details