JinYoung Kim

Hyojun Go, JinYoung Kim, Yunsung Lee et al. · cmu

Diffusion-based generative models have achieved remarkable success in various domains. It trains a shared model on denoising tasks that encompass different noise levels simultaneously, representing a form of multi-task learning (MTL). However, analyzing and improving diffusion models from an MTL perspective remains under-explored. In particular, MTL can sometimes lead to the well-known phenomenon of negative transfer, which results in the performance degradation of certain tasks due to conflicts between tasks. In this paper, we first aim to analyze diffusion training from an MTL standpoint, presenting two key observations: (O1) the task affinity between denoising tasks diminishes as the gap between noise levels widens, and (O2) negative transfer can arise even in diffusion training. Building upon these observations, we aim to enhance diffusion training by mitigating negative transfer. To achieve this, we propose leveraging existing MTL methods, but the presence of a huge number of denoising tasks makes this computationally expensive to calculate the necessary per-task loss or gradient. To address this challenge, we propose clustering the denoising tasks into small task clusters and applying MTL methods to them. Specifically, based on (O2), we employ interval clustering to enforce temporal proximity among denoising tasks within clusters. We show that interval clustering can be solved using dynamic programming, utilizing signal-to-noise ratio, timestep, and task affinity for clustering objectives. Through this, our approach addresses the issue of negative transfer in diffusion models by allowing for efficient computation of MTL methods. We validate the efficacy of proposed clustering and its integration with MTL methods through various experiments, demonstrating 1) improved generation quality and 2) faster training convergence of diffusion models.

Jiho Jang, Jinyoung Kim, Kyungjune Baek et al.

Reinforcement Learning from Human Feedback has emerged as a standard for aligning diffusion models. However, we identify a fundamental limitation in the standard DPO formulation because it relies on the Bradley-Terry model to aggregate diverse evaluation axes like aesthetic quality and semantic alignment into a single scalar reward. This aggregation creates a reward conflict where the model is forced to unlearn desirable features of a specific dimension if they appear in a globally non-preferred sample. To address this issue, we propose Multi Reward Conditional DPO (MCDPO). This method resolves reward conflicts by introducing a disentangled Bradley-Terry objective. MCDPO explicitly injects a preference outcome vector as a condition during training, which allows the model to learn the correct optimization direction for each reward axis independently within a single network. We further introduce dimensional reward dropout to ensure balanced optimization across dimensions. Extensive experiments on Stable Diffusion 1.5 and SDXL demonstrate that MCDPO achieves superior performance on benchmarks. Notably, our conditional framework enables dynamic and multiple-axis control at inference time using Classifier Free Guidance to amplify specific reward dimensions without additional training or external reward models.

Jongha Kim, Jihwan Park, Jinyoung Park et al.

Visual Relationship Detection (VRD) has seen significant advancements with Transformer-based architectures recently. However, we identify two key limitations in a conventional label assignment for training Transformer-based VRD models, which is a process of mapping a ground-truth (GT) to a prediction. Under the conventional assignment, an unspecialized query is trained since a query is expected to detect every relation, which makes it difficult for a query to specialize in specific relations. Furthermore, a query is also insufficiently trained since a GT is assigned only to a single prediction, therefore near-correct or even correct predictions are suppressed by being assigned no relation as a GT. To address these issues, we propose Groupwise Query Specialization and Quality-Aware Multi-Assignment (SpeaQ). Groupwise Query Specialization trains a specialized query by dividing queries and relations into disjoint groups and directing a query in a specific query group solely toward relations in the corresponding relation group. Quality-Aware Multi-Assignment further facilitates the training by assigning a GT to multiple predictions that are significantly close to a GT in terms of a subject, an object, and the relation in between. Experimental results and analyses show that SpeaQ effectively trains specialized queries, which better utilize the capacity of a model, resulting in consistent performance gains with zero additional inference cost across multiple VRD models and benchmarks. Code is available at https://github.com/mlvlab/SpeaQ.

Jungbae Park, Jinyoung Kim, Soonwoo Kwon et al.

Knowledge tracing and dropout prediction are crucial for online education to estimate students' knowledge states or to prevent dropout rates. While traditional systems interacting with students suffered from data sparsity and overfitting, recent sample-level contrastive learning helps to alleviate this issue. One major limitation of sample-level approaches is that they regard students' behavior interaction sequences as a bundle, so they often fail to encode temporal contexts and track their dynamic changes, making it hard to find optimal representations for knowledge tracing and dropout prediction. To apply temporal context within the sequence, this study introduces a novel student modeling framework, SAICL: \textbf{s}tudent modeling with \textbf{a}uxiliary \textbf{i}nteraction-level \textbf{c}ontrastive \textbf{l}earning. In detail, SAICL can utilize both proposed self-supervised/supervised interaction-level contrastive objectives: MilCPC (\textbf{M}ulti-\textbf{I}nteraction-\textbf{L}evel \textbf{C}ontrastive \textbf{P}redictive \textbf{C}oding) and SupCPC (\textbf{Sup}ervised \textbf{C}ontrastive \textbf{P}redictive \textbf{C}oding). While previous sample-level contrastive methods for student modeling are highly dependent on data augmentation methods, the SAICL is free of data augmentation while showing better performance in both self-supervised and supervised settings. By combining cross-entropy with contrastive objectives, the proposed SAICL achieved comparable knowledge tracing and dropout prediction performance with other state-of-art models without compromising inference costs.

Hyeonwoo Cho, Chanmin Park, Dong-Hee Kim et al.

Domain shift occurs when training (source) and test (target) data diverge in their distribution. Source-Free Domain Adaptation (SFDA) addresses this domain shift problem, aiming to adopt a trained model on the source domain to the target domain in a scenario where only a well-trained source model and unlabeled target data are available. In this scenario, handling false labels in the target domain is crucial because they negatively impact the model performance. To deal with this problem, we propose to update cluster prototypes (i.e., centroid of each sample cluster) and their structure in the target domain formulated by the source model in online manners. In the feature space, samples in different regions have different pseudo-label distribution characteristics affected by the cluster prototypes, and we adopt distinct training strategies for these samples by defining clean and noisy regions: we selectively train the target with clean pseudo-labels in the clean region, whereas we introduce mix-up inputs representing intermediate features between clean and noisy regions to increase the compactness of the cluster. We conducted extensive experiments on multiple datasets in online/offline SFDA settings, whose results demonstrate that our method, CNG-SFDA, achieves state-of-the-art for most cases. Code is available at https://github.com/hyeonwoocho7/CNG-SFDA.

Jinyoung Kim, Hyeongsoo Lim, Eunseo Seo et al.

Recent advances in large audio language models (LALMs) have enabled multilingual speech understanding. However, benchmarks for evaluating LALMs remain scarce for non-English languages, with Korean being one such underexplored case. In this paper, we introduce KoALa-Bench, a comprehensive benchmark for evaluating Korean speech understanding and speech faithfulness of LALMs. In particular, KoALa-Bench comprises six tasks. Four tasks evaluate fundamental speech understanding capabilities, including automatic speech recognition, speech translation, speech question answering, and speech instruction following, while the remaining two tasks evaluate speech faithfulness, motivated by our observation that several LALMs often fail to fully leverage the speech modality. Furthermore, to reflect Korea-specific knowledge, our benchmark incorporates listening questions from the Korean college scholastic ability test as well as content covering Korean cultural domains. We conduct extensive experiments across six models, including both white-box and black-box ones. Our benchmark, evaluation code, and leaderboard are publicly available at https://ksbench.github.io/Korean-Benchmark/.

Jinyoung Park, Jeehye Na, Jinyoung Kim et al.

Recent works have demonstrated the effectiveness of reinforcement learning (RL)-based post-training for enhancing the reasoning capabilities of large language models (LLMs). In particular, Group Relative Policy Optimization (GRPO) has shown impressive success using a PPO-style reinforcement algorithm with group-normalized rewards. However, the effectiveness of GRPO in Video Large Language Models (VideoLLMs) has still been less studyed. In this paper, we explore GRPO and identify two problems that deteriorate the effective learning: (1) reliance on safeguards, and (2) vanishing advantage. To mitigate these challenges, we propose DeepVideo-R1, a video large language model trained with Reg-GRPO (Regressive GRPO) and difficulty-aware data augmentation. Reg-GRPO reformulates the GRPO loss function into a regression task that directly predicts the advantage in GRPO, eliminating the need for safeguards such as the clipping and min functions. It directly aligns the model with advantages, providing guidance to prefer better ones. The difficulty-aware data augmentation strategy augments input prompts/videos to locate the difficulty of samples at solvable difficulty levels, enabling diverse reward signals. Our experimental results show that our approach significantly improves video reasoning performance across multiple benchmarks.

Dong-Hee Kim, Sungduk Cho, Hyeonwoo Cho et al.

In this work, we introduce Mask-JEPA, a self-supervised learning framework tailored for mask classification architectures (MCA), to overcome the traditional constraints associated with training segmentation models. Mask-JEPA combines a Joint Embedding Predictive Architecture with MCA to adeptly capture intricate semantics and precise object boundaries. Our approach addresses two critical challenges in self-supervised learning: 1) extracting comprehensive representations for universal image segmentation from a pixel decoder, and 2) effectively training the transformer decoder. The use of the transformer decoder as a predictor within the JEPA framework allows proficient training in universal image segmentation tasks. Through rigorous evaluations on datasets such as ADE20K, Cityscapes and COCO, Mask-JEPA demonstrates not only competitive results but also exceptional adaptability and robustness across various training scenarios. The architecture-agnostic nature of Mask-JEPA further underscores its versatility, allowing seamless adaptation to various mask classification family.

Jinyoung Kim, Dayoon Ko, Gunhee Kim

In the rapidly evolving landscape of language, resolving new linguistic expressions in continuously updating knowledge bases remains a formidable challenge. This challenge becomes critical in retrieval-augmented generation (RAG) with knowledge bases, as emerging expressions hinder the retrieval of relevant documents, leading to generator hallucinations. To address this issue, we introduce a novel task aimed at resolving emerging mentions to dynamic entities and present DynamicER benchmark. Our benchmark includes dynamic entity mention resolution and entity-centric knowledge-intensive QA task, evaluating entity linking and RAG model's adaptability to new expressions, respectively. We discovered that current entity linking models struggle to link these new expressions to entities. Therefore, we propose a temporal segmented clustering method with continual adaptation, effectively managing the temporal dynamics of evolving entities and emerging mentions. Extensive experiments demonstrate that our method outperforms existing baselines, enhancing RAG model performance on QA task with resolved mentions.

Jeonghyo Song, Kimin Yun, DaeUng Jo et al.

Effective Out-of-Distribution (OOD) detection is criti-cal for ensuring the reliability of semantic segmentation models, particularly in complex road environments where safety and accuracy are paramount. Despite recent advancements in large language models (LLMs), notably GPT-4, which significantly enhanced multimodal reasoning through Chain-of-Thought (CoT) prompting, the application of CoT-based visual reasoning for OOD semantic segmentation remains largely unexplored. In this paper, through extensive analyses of the road scene anomalies, we identify three challenging scenarios where current state-of-the-art OOD segmentation methods consistently struggle: (1) densely packed and overlapping objects, (2) distant scenes with small objects, and (3) large foreground-dominant objects. To address the presented challenges, we propose a novel CoT-based framework targeting OOD detection in road anomaly scenes. Our method leverages the extensive knowledge and reasoning capabilities of foundation models, such as GPT-4, to enhance OOD detection through improved image understanding and prompt-based reasoning aligned with observed problematic scene attributes. Extensive experiments show that our framework consistently outperforms state-of-the-art methods on both standard benchmarks and our newly defined challenging subset of the RoadAnomaly dataset, offering a robust and interpretable solution for OOD semantic segmentation in complex driving environments.

Dayoon Ko, Jinyoung Kim, Sohyeon Kim et al.

Dense retrievers encode texts into embeddings to efficiently retrieve relevant documents from large databases in response to user queries. However, real-world corpora continually evolve, leading to a shift from the original training distribution of the retriever. Without timely updates or retraining, indexing newly emerging documents can degrade retrieval performance for future queries. Thus, identifying when a dense retriever requires an update is critical for maintaining robust retrieval systems. In this paper, we propose a novel task of predicting whether a corpus is out-of-distribution (OOD) relative to a dense retriever before indexing. Addressing this task allows us to proactively manage retriever updates, preventing potential retrieval failures. We introduce GradNormIR, an unsupervised approach that leverages gradient norms to detect OOD corpora effectively. Experiments on the BEIR benchmark demonstrate that GradNormIR enables timely updates of dense retrievers in evolving document collections, significantly enhancing retrieval robustness and efficiency.

JinYoung Kim, DaeUng Jo, Kimin Yun et al.

Estimating uncertainty from deep neural networks is a widely used approach for detecting out-of-distribution (OoD) samples, which typically exhibit high predictive uncertainty. However, conventional methods such as Monte Carlo (MC) Dropout often focus solely on either model or data uncertainty, failing to align with the semantic objective of OoD detection. To address this, we propose the Free-Energy Posterior Network, a novel framework that jointly models distributional uncertainty and identifying OoD and misclassified regions using free energy. Our method introduces two key contributions: (1) a free-energy-based density estimator parameterized by a Beta distribution, which enables fine-grained uncertainty estimation near ambiguous or unseen regions; and (2) a loss integrated within a posterior network, allowing direct uncertainty estimation from learned parameters without requiring stochastic sampling. By integrating our approach with the residual prediction branch (RPL) framework, the proposed method goes beyond post-hoc energy thresholding and enables the network to learn OoD regions by leveraging the variance of the Beta distribution, resulting in a semantically meaningful and computationally efficient solution for uncertainty-aware segmentation. We validate the effectiveness of our method on challenging real-world benchmarks, including Fishyscapes, RoadAnomaly, and Segment-Me-If-You-Can.

Dayoon Ko, Jinyoung Kim, Hahyeon Choi et al.

In the real world, knowledge is constantly evolving, which can render existing knowledge-based datasets outdated. This unreliability highlights the critical need for continuous updates to ensure both accuracy and relevance in knowledge-intensive tasks. To address this, we propose GrowOVER-QA and GrowOVER-Dialogue, dynamic open-domain QA and dialogue benchmarks that undergo a continuous cycle of updates, keeping pace with the rapid evolution of knowledge. Our research indicates that retrieval-augmented language models (RaLMs) struggle with knowledge that has not been trained on or recently updated. Consequently, we introduce a novel retrieval-interactive language model framework, where the language model evaluates and reflects on its answers for further re-retrieval. Our exhaustive experiments demonstrate that our training-free framework significantly improves upon existing methods, performing comparably to or even surpassing continuously trained language models.

Jinyoung Kim, Remi Patriat, Jordan Kaplan et al.

Deep cerebellar nuclei are a key structure of the cerebellum that are involved in processing motor and sensory information. It is thus a crucial step to accurately segment deep cerebellar nuclei for the understanding of the cerebellum system and its utility in deep brain stimulation treatment. However, it is challenging to clearly visualize such small nuclei under standard clinical magnetic resonance imaging (MRI) protocols and therefore precise segmentation is not feasible. Recent advances in 7 Tesla (T) MRI technology and great potential of deep neural networks facilitate automatic patient-specific segmentation. In this paper, we propose a novel deep learning framework (referred to as DCN-Net) for fast, accurate, and robust patient-specific segmentation of deep cerebellar dentate and interposed nuclei on 7T diffusion MRI. DCN-Net effectively encodes contextual information on the patch images without consecutive pooling operations and adding complexity via proposed dilated dense blocks. During the end-to-end training, label probabilities of dentate and interposed nuclei are independently learned with a hybrid loss, handling highly imbalanced data. Finally, we utilize self-training strategies to cope with the problem of limited labeled data. To this end, auxiliary dentate and interposed nuclei labels are created on unlabeled data by using DCN-Net trained on manual labels. We validate the proposed framework using 7T B0 MRIs from 60 subjects. Experimental results demonstrate that DCN-Net provides better segmentation than atlas-based deep cerebellar nuclei segmentation tools and other state-of-the-art deep neural networks in terms of accuracy and consistency. We further prove the effectiveness of the proposed components within DCN-Net in dentate and interposed nuclei segmentation.

Seongmin Hwang, Gwanghuyn Yu, Cheolkon Jung et al.

Although deep convolutional neural networks (CNNs) have obtained outstanding performance in image superresolution (SR), their computational cost increases geometrically as CNN models get deeper and wider. Meanwhile, the features of intermediate layers are treated equally across the channel, thus hindering the representational capability of CNNs. In this paper, we propose an attention-aware linear depthwise network to address the problems for single image SR, named ALDNet. Specifically, linear depthwise convolution allows CNN-based SR models to preserve useful information for reconstructing a super-resolved image while reducing computational burden. Furthermore, we design an attention-aware branch that enhances the representation ability of depthwise convolution layers by making full use of depthwise filter interdependency. Experiments on publicly available benchmark datasets show that ALDNet achieves superior performance to traditional depthwise separable convolutions in terms of quantitative measurements and visual quality.

Reuben R Shamir, Yuval Duchin, Jinyoung Kim et al.

Objective: Overlapping measures are often utilized to quantify the similarity between two binary regions. However, modern segmentation algorithms output a probability or confidence map with continuous values in the zero-to-one interval. Moreover, these binary overlapping measures are biased to structure size. Addressing these challenges is the objective of this work. Methods: We extend the definition of the classical Dice coefficient (DC) overlap to facilitate the direct comparison of a ground truth binary image with a probabilistic map. We call the extended method continuous Dice coefficient (cDC) and show that 1) cDC is less or equal to 1 and cDC = 1 if-and-only-if the structures overlap is complete, and, 2) cDC is monotonically decreasing with the amount of overlap. We compare the classical DC and the cDC in a simulation of partial volume effects that incorporates segmentations of common targets for deep-brainstimulation. Lastly, we investigate the cDC for an automatic segmentation of the subthalamic-nucleus. Results: Partial volume effect simulation on thalamus (large structure) resulted with DC and cDC averages (SD) of 0.98 (0.006) and 0.99 (0.001), respectively. For subthalamic-nucleus (small structure) DC and cDC were 0.86 (0.025) and 0.97 (0.006), respectively. The DC and cDC for automatic STN segmentation were 0.66 and 0.80, respectively. Conclusion: The cDC is well defined for probabilistic segmentation, less biased to structure size and more robust to partial volume effects in comparison to DC. Significance: The proposed method facilitates a better evaluation of segmentation algorithms. As a better measurement tool, it opens the door for the development of better segmentation methods.

Seongmin Hwang, Gwanghyun Yu, Huy Toan Nguyen et al.

Thermal cameras show noisy images due to their limited thermal resolution, especially for the scenes of a low temperature difference. In order to deal with a noise problem, this paper proposes a novel neural network architecture with repeatable denoising inception-residual blocks(DnIRB) for noise learning. Each DnIRB has two sub-blocks with difference receptive fields and one shortcut connection to prevent a vanishing gradient problem. The proposed approach is tested for thermal images. The experimental results indicate that the proposed approach shows the best SQNR performance and reasonable processing time compared with state-of-the-art denoising methods.