Author name cluster

Chun-Yi Lee

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.

22 papers

2 author rows

NeurIPS Conference 2025 Conference Paper

Beyond Masked and Unmasked: Discrete Diffusion Models via Partial Masking

Chen-Hao (Lance) Chao
Wei-Fang Sun
Hanwen Liang
Chun-Yi Lee
Rahul Krishnan

Masked diffusion models (MDM) are powerful generative models for discrete data that generate samples by progressively unmasking tokens in a sequence. Each token can take one of two states: masked or unmasked. We observe that token sequences often remain unchanged between consecutive sampling steps; consequently, the model repeatedly processes identical inputs, leading to redundant computation. To address this inefficiency, we propose the Partial masking scheme (Prime), which augments MDM by allowing tokens to take intermediate states interpolated between the masked and unmasked states. This design enables the model to make predictions based on partially observed token information, and facilitates a fine-grained denoising process. We derive a variational training objective and introduce a simple architectural design to accommodate intermediate-state inputs. Our method demonstrates superior performance across a diverse set of generative modeling tasks. On text data, it achieves a perplexity of 15. 36 on OpenWebText, outperforming previous MDM (21. 52), autoregressive models (17. 54), and their hybrid variants (17. 58), without relying on an autoregressive formulation. On image data, it attains competitive FID scores of 3. 26 on CIFAR-10 and 6. 98 on ImageNet-32, comparable to leading continuous generative models.

PDF Details

NeurIPS Conference 2025 Conference Paper

EDELINE: Enhancing Memory in Diffusion-based World Models via Linear-Time Sequence Modeling

Jia-Hua Lee
Bor-Jiun Lin
Wei-Fang Sun
Chun-Yi Lee

World models represent a promising approach for training reinforcement learning agents with significantly improved sample efficiency. While most world model methods primarily rely on sequences of discrete latent variables to model environment dynamics, this compression often neglects critical visual details essential for reinforcement learning. Recent diffusion-based world models condition generation on a fixed context length of frames to predict the next observation, using separate recurrent neural networks to model rewards and termination signals. Although this architecture effectively enhances visual fidelity, the fixed context length approach inherently limits memory capacity. In this paper, we introduce EDELINE, a unified world model architecture that integrates state space models with diffusion models. Our approach outperforms existing baselines across visually challenging Atari 100k tasks, memory-demanding Crafter benchmark, and 3D first-person ViZDoom environments, demonstrating superior performance in all these diverse challenges. Code is available at https: //github. com/LJH-coding/EDELINE.

PDF Details

ICML Conference 2025 Conference Paper

Retraining-free Merging of Sparse MoE via Hierarchical Clustering

I-Chun Chen
Hsu-Shen Liu
Wei-Fang Sun
Chen-Hao Chao
Yen-Chang Hsu
Chun-Yi Lee

Sparse Mixture-of-Experts (SMoE) models represent a significant advancement in large language model (LLM) development through their efficient parameter utilization. These models achieve substantial performance improvements at reduced inference costs. However, the deployment of SMoE models faces constraints from extensive memory requirements of expert components in resource-limited environments. To address these limitations, this paper introduces Hierarchical Clustering for Sparsely activated Mixture of Experts (HC-SMoE), a task-agnostic expert merging framework for parameter reduction without retraining. HC-SMoE introduces a novel hierarchical clustering approach based on expert outputs to ensure merging robustness independent of routing decisions. The proposed output-based clustering method enables effective capture of functional relationships between experts for largescale architectures. We provide theoretical analysis and comprehensive evaluations across multiple zero-shot language tasks to demonstrate HC-SMoE’s effectiveness in state-of-the-art models including Qwen and Mixtral. The experimental results validate HC-SMoE’s superior performance and practical applicability for real-world deployments. Our implementation is available at https: //github. com/wazenmai/HC-SMoE.

Details

ICML Conference 2024 Conference Paper

Expert Proximity as Surrogate Rewards for Single Demonstration Imitation Learning

Chia-Cheng Chiang
Li-Cheng Lan
Wei-Fang Sun
Chien Feng
Cho-Jui Hsieh
Chun-Yi Lee

In this paper, we focus on single-demonstration imitation learning (IL), a practical approach for real-world applications where acquiring multiple expert demonstrations is costly or infeasible and the ground truth reward function is not available. In contrast to typical IL settings with multiple demonstrations, single-demonstration IL involves an agent having access to only one expert trajectory. We highlight the issue of sparse reward signals in this setting and propose to mitigate this issue through our proposed Transition Discriminator-based IL (TDIL) method. TDIL is an IRL method designed to address reward sparsity by introducing a denser surrogate reward function that considers environmental dynamics. This surrogate reward function encourages the agent to navigate towards states that are proximal to expert states. In practice, TDIL trains a transition discriminator to differentiate between valid and non-valid transitions in a given environment to compute the surrogate rewards. The experiments demonstrate that TDIL outperforms existing IL approaches and achieves expert-level performance in the single-demonstration IL setting across five widely adopted MuJoCo benchmarks as well as the "Adroit Door" robotic environment.

Details

ICML Conference 2024 Conference Paper

HGAP: Boosting Permutation Invariant and Permutation Equivariant in Multi-Agent Reinforcement Learning via Graph Attention Network

Bor-Jiun Lin
Chun-Yi Lee

Graph representation has gained widespread application across various machine learning domains, attributed to its ability to discern correlations among input nodes. In the realm of Multi- agent Reinforcement Learning (MARL), agents are tasked with observing other entities within their environment to determine their behavior. Conventional MARL methodologies often suffer from training difficulties if Permutation Invariant (PI) and Permutation Equivariant (PE) properties are not considered during training. The adoption of graph representation offers a solution to these challenges by conceptualizing observed entities as a graph. In this context, we introduce the Hyper Graphical Attention Policy (HGAP) Network, which employs a graph attention mechanism to fulfill the PI and PE properties, while also understanding inter-entity interactions for decision-making. HGAP is assessed across various MARL benchmarks to confirm its effectiveness and efficiency. In addition, a series of ablation studies are provided to demonstrate its adaptability, transferability, and the capability to alleviate the complexities introduced by the POMDP constraint.

Details

IROS Conference 2024 Conference Paper

Language-Guided Pattern Formation for Swarm Robotics with Multi-Agent Reinforcement Learning

Hsu-Shen Liu
So Kuroki
Tadashi Kozuno
Wei-Fang Sun
Chun-Yi Lee

This paper explores leveraging the vast knowledge encoded in Large Language Models (LLMs) to tackle pattern formation challenges for swarm robotics systems. A new framework, named LGPF (Language-Guided Pattern Formation), is proposed to address these challenges. The framework breaks down the pattern formation into two key components: pattern synthesis and swarm robotics control. For the former, this study utilizes the exceptional few-shot generalizability of LLMs to translate high-level natural language descriptions into the desired spatial pattern coordinates. This approach allows for overcoming previous limitations in representing and designing complex patterns. The framework further employs a centralized training with decentralized execution (CTDE) based multiagent reinforcement learning (MARL) approach to control the swarm robots in forming the specified pattern while avoiding collisions. The decentralized policies learned with the CTDE-based MARL algorithm consider coordination between robots without direct communication under a partially observable setup. To validate the effectiveness of our framework, we perform extensive experiments in both simulation and real-world environments. These experiments validate LGPF’s effectiveness in accurately and safely forming diverse user-specified patterns.

Details

NeurIPS Conference 2024 Conference Paper

Maximum Entropy Reinforcement Learning via Energy-Based Normalizing Flow

Chen-Hao Chao
Chien Feng
Wei-Fang Sun
Cheng-Kuang Lee
Simon See
Chun-Yi Lee

Existing Maximum-Entropy (MaxEnt) Reinforcement Learning (RL) methods for continuous action spaces are typically formulated based on actor-critic frameworks and optimized through alternating steps of policy evaluation and policy improvement. In the policy evaluation steps, the critic is updated to capture the soft Q-function. In the policy improvement steps, the actor is adjusted in accordance with the updated soft Q-function. In this paper, we introduce a new MaxEnt RL framework modeled using Energy-Based Normalizing Flows (EBFlow). This framework integrates the policy evaluation steps and the policy improvement steps, resulting in a single objective training process. Our method enables the calculation of the soft value function used in the policy evaluation target without Monte Carlo approximation. Moreover, this design supports the modeling of multi-modal action distributions while facilitating efficient action sampling. To evaluate the performance of our method, we conducted experiments on the MuJoCo benchmark suite and a number of high-dimensional robotic tasks simulated by Omniverse Isaac Gym. The evaluation results demonstrate that our method achieves superior performance compared to widely-adopted representative baselines.

PDF Details DOI

IROS Conference 2024 Conference Paper

Precise Pick-and-Place using Score-Based Diffusion Networks

Shih-Wei Guo
Tsu-Ching Hsiao
Yu-Lun Liu 0001
Chun-Yi Lee

In this paper, we propose a novel coarse-to-fine continuous pose diffusion method to enhance the precision of pick-and-place operations within robotic manipulation tasks. Leveraging the capabilities of diffusion networks, we facilitate the accurate perception of object poses. This accurate perception enhances both pick-and-place success rates and overall manipulation precision. Our methodology utilizes a top-down RGB image projected from an RGB-D camera and adopts a coarse-to-fine architecture. This architecture enables efficient learning of coarse and fine models. A distinguishing feature of our approach is its focus on continuous pose estimation, which enables more precise object manipulation, particularly concerning rotational angles. In addition, we employ pose and color augmentation techniques to enable effective training with limited data. Through extensive experiments in simulated and real-world scenarios, as well as an ablation study, we comprehensively evaluate our proposed methodology. Taken together, the findings validate its effectiveness in achieving high-precision pick-and-place tasks.

Details

IROS Conference 2024 Conference Paper

Visual Forecasting as a Mid-level Representation for Avoidance

Hsuan-Kung Yang
Tsung-Chih Chiang
Ting-Ru Liu
Chun-Wei Huang
Jou-Min Liu
Chun-Yi Lee

The challenge of navigation in environments with dynamic objects continues to be a central issue in the study of autonomous agents. While predictive methods hold promise, their reliance on precise state information makes them less practical for real-world implementation. This study presents visual forecasting as an innovative alternative. By introducing intuitive visual cues, this approach projects the future trajectories of dynamic objects to improve agent perception and enable anticipatory actions. Our research explores two distinct strategies for conveying predictive information through visual forecasting: (1) sequences of bounding boxes, and (2) augmented paths. To validate the proposed visual forecasting strategies, we initiate evaluations in simulated environments using the Unity engine and then extend these evaluations to real-world scenarios to assess both practicality and effectiveness. The results confirm the viability of visual forecasting as a promising solution for navigation and obstacle avoidance in dynamic environments.

Details

JMLR Journal 2023 Journal Article

A Unified Framework for Factorizing Distributional Value Functions for Multi-Agent Reinforcement Learning

Wei-Fang Sun
Cheng-Kuang Lee
Simon See
Chun-Yi Lee

In fully cooperative multi-agent reinforcement learning (MARL) settings, environments are highly stochastic due to the partial observability of each agent and the continuously changing policies of other agents. To address the above issues, we proposed a unified framework, called DFAC, for integrating distributional RL with value function factorization methods. This framework generalizes expected value function factorization methods to enable the factorization of return distributions. To validate DFAC, we first demonstrate its ability to factorize the value functions of a simple matrix game with stochastic rewards. Then, we perform experiments on all Super Hard maps of the StarCraft Multi-Agent Challenge and six self-designed Ultra Hard maps, showing that DFAC is able to outperform a number of baselines. [abs] [ pdf ][ bib ] [ code ] &copy JMLR 2023. ( edit, beta )

PDF Details

ICML Conference 2023 Conference Paper

On Investigating the Conservative Property of Score-Based Generative Models

Chen-Hao Chao
Wei-Fang Sun
Bo-Wun Cheng
Chun-Yi Lee

Existing Score-Based Models (SBMs) can be categorized into constrained SBMs (CSBMs) or unconstrained SBMs (USBMs) according to their parameterization approaches. CSBMs model probability density functions as Boltzmann distributions, and assign their predictions as the negative gradients of some scalar-valued energy functions. On the other hand, USBMs employ flexible architectures capable of directly estimating scores without the need to explicitly model energy functions. In this paper, we demonstrate that the architectural constraints of CSBMs may limit their modeling ability. In addition, we show that USBMs’ inability to preserve the property of conservativeness may lead to degraded performance in practice. To address the above issues, we propose Quasi-Conservative Score-Based Models (QCSBMs) for keeping the advantages of both CSBMs and USBMs. Our theoretical derivations demonstrate that the training objective of QCSBMs can be efficiently integrated into the training processes by leveraging the Hutchinson’s trace estimator. In addition, our experimental results on the CIFAR-10, CIFAR-100, ImageNet, and SVHN datasets validate the effectiveness of QCSBMs. Finally, we justify the advantage of QCSBMs using an example of a one-layered autoencoder.

Details

NeurIPS Conference 2023 Conference Paper

Training Energy-Based Normalizing Flow with Score-Matching Objectives

Chen-Hao Chao
Wei-Fang Sun
Yen-Chang Hsu
Zsolt Kira
Chun-Yi Lee

In this paper, we establish a connection between the parameterization of flow-based and energy-based generative models, and present a new flow-based modeling approach called energy-based normalizing flow (EBFlow). We demonstrate that by optimizing EBFlow with score-matching objectives, the computation of Jacobian determinants for linear transformations can be entirely bypassed. This feature enables the use of arbitrary linear layers in the construction of flow-based models without increasing the computational time complexity of each training iteration from $\mathcal{O}(D^2L)$ to $\mathcal{O}(D^3L)$ for an $L$-layered model that accepts $D$-dimensional inputs. This makes the training of EBFlow more efficient than the commonly-adopted maximum likelihood training method. In addition to the reduction in runtime, we enhance the training stability and empirical performance of EBFlow through a number of techniques developed based on our analysis of the score-matching methods. The experimental results demonstrate that our approach achieves a significant speedup compared to maximum likelihood estimation while outperforming prior methods with a noticeable margin in terms of negative log-likelihood (NLL).

PDF Details

ICLR Conference 2022 Conference Paper

Denoising Likelihood Score Matching for Conditional Score-based Data Generation

Chen-Hao Chao
Wei-Fang Sun
Bo-Wun Cheng
Yi-Chen Lo
Chia-Che Chang
Yu-Lun Liu 0001
Yu-Lin Chang
Chia-Ping Chen

Many existing conditional score-based data generation methods utilize Bayes' theorem to decompose the gradients of a log posterior density into a mixture of scores. These methods facilitate the training procedure of conditional score models, as a mixture of scores can be separately estimated using a score model and a classifier. However, our analysis indicates that the training objectives for the classifier in these methods may lead to a serious score mismatch issue, which corresponds to the situation that the estimated scores deviate from the true ones. Such an issue causes the samples to be misled by the deviated scores during the diffusion process, resulting in a degraded sampling quality. To resolve it, we theoretically formulate a novel training objective, called Denoising Likelihood Score Matching (DLSM) loss, for the classifier to match the gradients of the true log likelihood density. Our experimental evidences show that the proposed method outperforms the previous methods on both Cifar-10 and Cifar-100 benchmarks noticeably in terms of several key evaluation metrics. We thus conclude that, by adopting DLSM, the conditional scores can be accurately modeled, and the effect of the score mismatch issue is alleviated.

Details

IROS Conference 2022 Conference Paper

Investigation of Factorized Optical Flows as Mid-Level Representations

Hsuan-Kung Yang
Tsu-Ching Hsiao
Ting-Hsuan Liao
Hsu-Shen Liu
Li-Yuan Tsao
Tzu-Wen Wang
Shan-Ya Yang
Yu-Wen Chen

In this paper, we introduce a new concept of incorporating factorized flow maps as mid-level representations, for bridging the perception and the control modules in modular learning based robotic frameworks. To investigate the advantages of factorized flow maps and examine their interplay with the other types of mid-level representations, we further develop a configurable framework, along with four different environments that contain both static and dynamic objects, for analyzing the impacts of factorized optical flow maps on the performance of deep reinforcement learning agents. Based on this framework, we report our experimental results on various scenarios, and offer a set of analyses to justify our hypothesis. Finally, we validate flow factorization in real world scenarios.

Details

AAMAS Conference 2021 Conference Paper

A Distributional Perspective on Value Function Factorization Methods for Multi-Agent Reinforcement Learning

Wei-Fang Sun
Cheng-Kuang Lee
Chun-Yi Lee

Distributional reinforcement learning (RL) provides beneficial impacts for the single-agent domain. However, distributional RL methods are not directly compatible with value function factorization methods for multi-agent reinforcement learning. This work provides a distributional perspective on value function factorization, offering a solution for bridging the gap between distributional RL and value function factorization methods.

PDF

ICML Conference 2021 Conference Paper

DFAC Framework: Factorizing the Value Function via Quantile Mixture for Multi-Agent Distributional Q-Learning

Wei-Fang Sun
Cheng-Kuang Lee
Chun-Yi Lee

In fully cooperative multi-agent reinforcement learning (MARL) settings, the environments are highly stochastic due to the partial observability of each agent and the continuously changing policies of the other agents. To address the above issues, we integrate distributional RL and value function factorization methods by proposing a Distributional Value Function Factorization (DFAC) framework to generalize expected value function factorization methods to their distributional variants. DFAC extends the individual utility functions from deterministic variables to random variables, and models the quantile function of the total return as a quantile mixture. To validate DFAC, we demonstrate DFAC’s ability to factorize a simple two-step matrix game with stochastic rewards and perform experiments on all Super Hard tasks of StarCraft Multi-Agent Challenge, showing that DFAC is able to outperform expected value function factorization baselines.

Details

ICRA Conference 2021 Conference Paper

Reducing the Deployment-Time Inference Control Costs of Deep Reinforcement Learning Agents via an Asymmetric Architecture

Chin-Jui Chang
Yu-Wei Chu
Chao-Hsien Ting
Hao-Kang Liu
Zhang-Wei Hong
Chun-Yi Lee

Deep reinforcement learning (DRL) has been demonstrated to provide promising results in several challenging decision making and control tasks. However, the required inference costs of deep neural networks (DNNs) could prevent DRL from being applied to mobile robots which cannot afford high energy-consuming computations. To enable DRL methods to be affordable in such energy-limited platforms, we propose an asymmetric architecture that reduces the overall inference costs via switching between a computationally expensive policy and an economic one. The experimental results evaluated on a number of representative benchmark suites for robotic control tasks demonstrate that our method is able to reduce the inference costs while retaining the agent’s overall performance.

Details

IROS Conference 2020 Conference Paper

Dynamic Attention-based Visual Odometry

Xin-Yu Kuo
Chien Liu
Kai-Chen Lin
Evan Luo
Yu-Wen Chen
Chun-Yi Lee

This paper proposes a dynamic attention-based visual odometry framework (DAVO), a learning-based VO method, for estimating the ego-motion of a monocular camera. DAVO dynamically adjusts the attention weights on different semantic categories for different motion scenarios based on optical flow maps. These weighted semantic categories can then be used to generate attention maps that highlight the relative importance of different semantic regions in input frames for pose estimation. In order to examine the proposed DAVO, we perform a number of experiments on the KITTI Visual Odometry and SLAM benchmark suite to quantitatively and qualitatively inspect the impacts of the dynamically adjusted weights on the accuracy of the evaluated trajectories. Moreover, we design a set of ablation analyses to justify each of our design choices, and validate the effectiveness as well as the advantages of DAVO. Our experiments on the KITTI dataset shows that the proposed DAVO framework does provide satisfactory performance in ego-motion estimation, and is able deliver competitive performance when compared to the contemporary VO methods.

Details

IJCAI Conference 2020 Conference Paper

Flow-based Intrinsic Curiosity Module

Hsuan-Kung Yang
Po-Han Chiang
Min-Fong Hong
Chun-Yi Lee

In this paper, we focus on a prediction-based novelty estimation strategy upon the deep reinforcement learning (DRL) framework, and present a flow-based intrinsic curiosity module (FICM) to exploit the prediction errors from optical flow estimation as exploration bonuses. We propose the concept of leveraging motion features captured between consecutive observations to evaluate the novelty of observations in an environment. FICM encourages a DRL agent to explore observations with unfamiliar motion features, and requires only two consecutive frames to obtain sufficient information when estimating the novelty. We evaluate our method and compare it with a number of existing methods on multiple benchmark environments, including Atari games, Super Mario Bros. , and ViZDoom. We demonstrate that FICM is favorable to tasks or environments featuring moving objects, which allow FICM to utilize the motion features between consecutive observations. We further ablatively analyze the encoding efficiency of FICM, and discuss its applicable domains comprehensively. See here for our codes and demo videos.

PDF Details DOI

AAMAS Conference 2018 Conference Paper

A Deep Policy Inference Q-Network for Multi-Agent Systems

Zhang-Wei Hong
Shih-Yang Su
Tzu-Yun Shann
Yi-Hsiang Chang
Chun-Yi Lee

We present DPIQN, a deep policy inference Q-network that targets multi-agent systems composed of controllable agents, collaborators, and opponents that interact with each other. We focus on one challenging issue in such systems—modeling agents with varying strategies—and propose to employ “policy features” learned from raw observations (e. g. , raw images) of collaborators and opponents by inferring their policies. DPIQN incorporates the learned policy features as a hidden vector into its own deep Q-network (DQN), such that it is able to predict better Q values for the controllable agents than the state-of-the-art deep reinforcement learning models. We further propose an enhanced version of DPIQN, called deep recurrent policy inference Q-network (DRPIQN), for handling partial observability. Both DPIQN and DRPIQN are trained by an adaptive training procedure, which adjusts the network’s attention to learn the policy features and its own Q-values at different phases of the training process. We present a comprehensive analysis of DPIQN and DRPIQN, and highlight their effectiveness and generalizability in various multi-agent settings. Our models are evaluated in a classic soccer game involving both competitive and collaborative scenarios. Experimental results performed on 1 vs. 1 and 2 vs. 2 games show that DPIQN and DRPIQN demonstrate superior performance to the baseline DQN and deep recurrent Q-network (DRQN) models. We also explore scenarios in which collaborators or opponents dynamically change their policies, and show that DPIQN and DRPIQN do lead to better overall performance in terms of stability and mean scores.

PDF

NeurIPS Conference 2018 Conference Paper

Diversity-Driven Exploration Strategy for Deep Reinforcement Learning

Zhang-Wei Hong
Tzu-Yun Shann
Shih-Yang Su
Yi-Hsiang Chang
Tsu-Jui Fu
Chun-Yi Lee

Efficient exploration remains a challenging research problem in reinforcement learning, especially when an environment contains large state spaces, deceptive local optima, or sparse rewards. To tackle this problem, we present a diversity-driven approach for exploration, which can be easily combined with both off- and on-policy reinforcement learning algorithms. We show that by simply adding a distance measure to the loss function, the proposed methodology significantly enhances an agent's exploratory behaviors, and thus preventing the policy from being trapped in local optima. We further propose an adaptive scaling method for stabilizing the learning process. We demonstrate the effectiveness of our method in huge 2D gridworlds and a variety of benchmark environments, including Atari 2600 and MuJoCo. Experimental results show that our method outperforms baseline approaches in most tasks in terms of mean scores and exploration efficiency.

PDF Details

IJCAI Conference 2018 Conference Paper

Virtual-to-Real: Learning to Control in Visual Semantic Segmentation

Zhang-Wei Hong
Yu-Ming Chen
Hsuan-Kung Yang
Shih-Yang Su
Tzu-Yun Shann
Yi-Hsiang Chang
Brian Hsi-Lin Ho
Chih-Chieh Tu

Collecting training data from the physical world is usually time-consuming and even dangerous for fragile robots, and thus, recent advances in robot learning advocate the use of simulators as the training platform. Unfortunately, the reality gap between synthetic and real visual data prohibits direct migration of the models trained in virtual worlds to the real world. This paper proposes a modular architecture for tackling the virtual-to-real problem. The proposed architecture separates the learning model into a perception module and a control policy module, and uses semantic image segmentation as the meta representation for relating these two modules. The perception module translates the perceived RGB image to semantic image segmentation. The control policy module is implemented as a deep reinforcement learning agent, which performs actions based on the translated image segmentation. Our architecture is evaluated in an obstacle avoidance task and a target following task. Experimental results show that our architecture significantly outperforms all of the baseline methods in both virtual and real environments, and demonstrates a faster learning curve than them. We also present a detailed analysis for a variety of variant configurations, and validate the transferability of our modular architecture.

PDF Details