Jiyoung Lee

Jungin Park, Jiyoung Lee, Kwanghoon Sohn

This study introduces an intriguing phenomenon in Video LLMs: rather than merely translating frames into textual embeddings, Video LLMs establish a continuous manifold, token interface, allowing visual tokens to operate as standalone entities within the architecture. Exploiting this discovery, we propose V-LynX, a scalable framework that integrates novel modalities into Video LLMs by repurposing the internalized interface. Departing from conventional paradigms that necessitate heavy modality-specific encoders or paired supervision, V-LynX employs a lightweight auxiliary pathway in parallel with the frozen vision encoder. Our method integrates new sensory inputs with intrinsic video priors by aligning both attention responses and statistical distributions using unpaired unimodal data sets. This ensures manifold compatibility while preserving the integrity of the Video LLMs. Extensive benchmarks demonstrate that V-LynX achieves SOTA and efficiency across audio-visual QA, 3D reasoning, high-frame-rate, and multi-view video understanding. The code is available at https://github.com/park-jungin/lynx.

Jiyoung Lee, Seungho Kim, Seunghyun Won et al. · nvidia, utoronto

AI alignment refers to models acting towards human-intended goals, preferences, or ethical principles. Given that most large-scale deep learning models act as black boxes and cannot be manually controlled, analyzing the similarity between models and humans can be a proxy measure for ensuring AI safety. In this paper, we focus on the models' visual perception alignment with humans, further referred to as AI-human visual alignment. Specifically, we propose a new dataset for measuring AI-human visual alignment in terms of image classification, a fundamental task in machine perception. In order to evaluate AI-human visual alignment, a dataset should encompass samples with various scenarios that may arise in the real world and have gold human perception labels. Our dataset consists of three groups of samples, namely Must-Act (i.e., Must-Classify), Must-Abstain, and Uncertain, based on the quantity and clarity of visual information in an image and further divided into eight categories. All samples have a gold human perception label; even Uncertain (severely blurry) sample labels were obtained via crowd-sourcing. The validity of our dataset is verified by sampling theory, statistical theories related to survey design, and experts in the related fields. Using our dataset, we analyze the visual alignment and reliability of five popular visual perception models and seven abstention methods. Our code and data is available at https://github.com/jiyounglee-0523/VisAlign.

Hanjae Kim, Jiyoung Lee, Seongheon Park et al.

Compositional zero-shot learning (CZSL) aims to recognize unseen compositions with prior knowledge of known primitives (attribute and object). Previous works for CZSL often suffer from grasping the contextuality between attribute and object, as well as the discriminability of visual features, and the long-tailed distribution of real-world compositional data. We propose a simple and scalable framework called Composition Transformer (CoT) to address these issues. CoT employs object and attribute experts in distinctive manners to generate representative embeddings, using the visual network hierarchically. The object expert extracts representative object embeddings from the final layer in a bottom-up manner, while the attribute expert makes attribute embeddings in a top-down manner with a proposed object-guided attention module that models contextuality explicitly. To remedy biased prediction caused by imbalanced data distribution, we develop a simple minority attribute augmentation (MAA) that synthesizes virtual samples by mixing two images and oversampling minority attribute classes. Our method achieves SoTA performance on several benchmarks, including MIT-States, C-GQA, and VAW-CZSL. We also demonstrate the effectiveness of CoT in improving visual discrimination and addressing the model bias from the imbalanced data distribution. The code is available at https://github.com/HanjaeKim98/CoT.

Jungin Park, Jiyoung Lee, Kwanghoon Sohn

Model merging aims to consolidate multiple task-specific models fine-tuned on different datasets into a unified architecture that performs cross-domain proficiency. Current data-free model merging methods often struggle to scale as they rely on simple parameter-level heuristics that ignore inter-layer dependencies and non-uniform distribution of expertise. This work proposes SA-Merging, which is built upon connectivity-based saliency formulations from structural pruning (e.g., SynFlow) and extends them to the data-free model merging setting. We define a saliency score over task vectors relative to a shared base model, and further introduce merge-aware modulation that incorporates agreement across experts to mitigate task interference. Based on this formulation, an iterative saliency-aware merging procedure progressively removes non-informative updates while preserving end-to-end connectivity. Furthermore, we extend SA-Merging to introduce rank-wise saliency decomposition for LoRAs without compromising their structural integrity. Extensive experiments on vision and language tasks demonstrate the effectiveness of our saliency-based approach, further reducing the gap between data-free and test-time adaptation methods.

Gyeongsik Moon, Hongsuk Choi, Sanghyuk Chun et al.

Recovering 3D human mesh in the wild is greatly challenging as in-the-wild (ITW) datasets provide only 2D pose ground truths (GTs). Recently, 3D pseudo-GTs have been widely used to train 3D human mesh estimation networks as the 3D pseudo-GTs enable 3D mesh supervision when training the networks on ITW datasets. However, despite the great potential of the 3D pseudo-GTs, there has been no extensive analysis that investigates which factors are important to make more beneficial 3D pseudo-GTs. In this paper, we provide three recipes to obtain highly beneficial 3D pseudo-GTs of ITW datasets. The main challenge is that only 2D-based weak supervision is allowed when obtaining the 3D pseudo-GTs. Each of our three recipes addresses the challenge in each aspect: depth ambiguity, sub-optimality of weak supervision, and implausible articulation. Experimental results show that simply re-training state-of-the-art networks with our new 3D pseudo-GTs elevates their performance to the next level without bells and whistles. The 3D pseudo-GT is publicly available in https://github.com/mks0601/NeuralAnnot_RELEASE.

Junyoung Seo, Wooseok Jang, Min-Seop Kwak et al. · nvidia, utoronto

Text-to-3D generation has shown rapid progress in recent days with the advent of score distillation, a methodology of using pretrained text-to-2D diffusion models to optimize neural radiance field (NeRF) in the zero-shot setting. However, the lack of 3D awareness in the 2D diffusion models destabilizes score distillation-based methods from reconstructing a plausible 3D scene. To address this issue, we propose 3DFuse, a novel framework that incorporates 3D awareness into pretrained 2D diffusion models, enhancing the robustness and 3D consistency of score distillation-based methods. We realize this by first constructing a coarse 3D structure of a given text prompt and then utilizing projected, view-specific depth map as a condition for the diffusion model. Additionally, we introduce a training strategy that enables the 2D diffusion model learns to handle the errors and sparsity within the coarse 3D structure for robust generation, as well as a method for ensuring semantic consistency throughout all viewpoints of the scene. Our framework surpasses the limitations of prior arts, and has significant implications for 3D consistent generation of 2D diffusion models.

Junyoung Seo, Gyuseong Lee, Seokju Cho et al. · nvidia, utoronto

We present a novel method for exemplar-based image translation, called matching interleaved diffusion models (MIDMs). Most existing methods for this task were formulated as GAN-based matching-then-generation framework. However, in this framework, matching errors induced by the difficulty of semantic matching across cross-domain, e.g., sketch and photo, can be easily propagated to the generation step, which in turn leads to degenerated results. Motivated by the recent success of diffusion models overcoming the shortcomings of GANs, we incorporate the diffusion models to overcome these limitations. Specifically, we formulate a diffusion-based matching-and-generation framework that interleaves cross-domain matching and diffusion steps in the latent space by iteratively feeding the intermediate warp into the noising process and denoising it to generate a translated image. In addition, to improve the reliability of the diffusion process, we design a confidence-aware process using cycle-consistency to consider only confident regions during translation. Experimental results show that our MIDMs generate more plausible images than state-of-the-art methods.

Jin Kim, Jiyoung Lee, Jungin Park et al.

The rise of deep neural networks has led to several breakthroughs for semantic segmentation. In spite of this, a model trained on source domain often fails to work properly in new challenging domains, that is directly concerned with the generalization capability of the model. In this paper, we present a novel memory-guided domain generalization method for semantic segmentation based on meta-learning framework. Especially, our method abstracts the conceptual knowledge of semantic classes into categorical memory which is constant beyond the domains. Upon the meta-learning concept, we repeatedly train memory-guided networks and simulate virtual test to 1) learn how to memorize a domain-agnostic and distinct information of classes and 2) offer an externally settled memory as a class-guidance to reduce the ambiguity of representation in the test data of arbitrary unseen domain. To this end, we also propose memory divergence and feature cohesion losses, which encourage to learn memory reading and update processes for category-aware domain generalization. Extensive experiments for semantic segmentation demonstrate the superior generalization capability of our method over state-of-the-art works on various benchmarks.

Jungin Park, Jiyoung Lee, Kwanghoon Sohn

In this paper, we efficiently transfer the surpassing representation power of the vision foundation models, such as ViT and Swin, for video understanding with only a few trainable parameters. Previous adaptation methods have simultaneously considered spatial and temporal modeling with a unified learnable module but still suffered from fully leveraging the representative capabilities of image transformers. We argue that the popular dual-path (two-stream) architecture in video models can mitigate this problem. We propose a novel DualPath adaptation separated into spatial and temporal adaptation paths, where a lightweight bottleneck adapter is employed in each transformer block. Especially for temporal dynamic modeling, we incorporate consecutive frames into a grid-like frameset to precisely imitate vision transformers' capability that extrapolates relationships between tokens. In addition, we extensively investigate the multiple baselines from a unified perspective in video understanding and compare them with DualPath. Experimental results on four action recognition benchmarks prove that pretrained image transformers with DualPath can be effectively generalized beyond the data domain.

Jungin Park, Jiyoung Lee, Ig-Jae Kim et al.

This paper presents Probabilistic Video Contrastive Learning, a self-supervised representation learning method that bridges contrastive learning with probabilistic representation. We hypothesize that the clips composing the video have different distributions in short-term duration, but can represent the complicated and sophisticated video distribution through combination in a common embedding space. Thus, the proposed method represents video clips as normal distributions and combines them into a Mixture of Gaussians to model the whole video distribution. By sampling embeddings from the whole video distribution, we can circumvent the careful sampling strategy or transformations to generate augmented views of the clips, unlike previous deterministic methods that have mainly focused on such sample generation strategies for contrastive learning. We further propose a stochastic contrastive loss to learn proper video distributions and handle the inherent uncertainty from the nature of the raw video. Experimental results verify that our probabilistic embedding stands as a state-of-the-art video representation learning for action recognition and video retrieval on the most popular benchmarks, including UCF101 and HMDB51.

Dahye Kim, Jungin Park, Jiyoung Lee et al.

Given an untrimmed video and a language query depicting a specific temporal moment in the video, video grounding aims to localize the time interval by understanding the text and video simultaneously. One of the most challenging issues is an extremely time- and cost-consuming annotation collection, including video captions in a natural language form and their corresponding temporal regions. In this paper, we present a simple yet novel training framework for video grounding in the zero-shot setting, which learns a network with only video data without any annotation. Inspired by the recent language-free paradigm, i.e. training without language data, we train the network without compelling the generation of fake (pseudo) text queries into a natural language form. Specifically, we propose a method for learning a video grounding model by selecting a temporal interval as a hypothetical correct answer and considering the visual feature selected by our method in the interval as a language feature, with the help of the well-aligned visual-language space of CLIP. Extensive experiments demonstrate the prominence of our language-free training framework, outperforming the existing zero-shot video grounding method and even several weakly-supervised approaches with large margins on two standard datasets.

Yunji Kim, Jiyoung Lee, Jin-Hwa Kim et al.

Existing text-to-image diffusion models struggle to synthesize realistic images given dense captions, where each text prompt provides a detailed description for a specific image region. To address this, we propose DenseDiffusion, a training-free method that adapts a pre-trained text-to-image model to handle such dense captions while offering control over the scene layout. We first analyze the relationship between generated images' layouts and the pre-trained model's intermediate attention maps. Next, we develop an attention modulation method that guides objects to appear in specific regions according to layout guidance. Without requiring additional fine-tuning or datasets, we improve image generation performance given dense captions regarding both automatic and human evaluation scores. In addition, we achieve similar-quality visual results with models specifically trained with layout conditions.

Jiyoung Lee, Joon Son Chung, Soo-Whan Chung

The goal of this work is zero-shot text-to-speech synthesis, with speaking styles and voices learnt from facial characteristics. Inspired by the natural fact that people can imagine the voice of someone when they look at his or her face, we introduce a face-styled diffusion text-to-speech (TTS) model within a unified framework learnt from visible attributes, called Face-TTS. This is the first time that face images are used as a condition to train a TTS model. We jointly train cross-model biometrics and TTS models to preserve speaker identity between face images and generated speech segments. We also propose a speaker feature binding loss to enforce the similarity of the generated and the ground truth speech segments in speaker embedding space. Since the biometric information is extracted directly from the face image, our method does not require extra fine-tuning steps to generate speech from unseen and unheard speakers. We train and evaluate the model on the LRS3 dataset, an in-the-wild audio-visual corpus containing background noise and diverse speaking styles. The project page is https://facetts.github.io.

Soohyun Kim, Junho Kim, Taekyung Kim et al. · nvidia, utoronto

In this paper, we tackle the challenging task of Panoramic Image-to-Image translation (Pano-I2I) for the first time. This task is difficult due to the geometric distortion of panoramic images and the lack of a panoramic image dataset with diverse conditions, like weather or time. To address these challenges, we propose a panoramic distortion-aware I2I model that preserves the structure of the panoramic images while consistently translating their global style referenced from a pinhole image. To mitigate the distortion issue in naive 360 panorama translation, we adopt spherical positional embedding to our transformer encoders, introduce a distortion-free discriminator, and apply sphere-based rotation for augmentation and its ensemble. We also design a content encoder and a style encoder to be deformation-aware to deal with a large domain gap between panoramas and pinhole images, enabling us to work on diverse conditions of pinhole images. In addition, considering the large discrepancy between panoramas and pinhole images, our framework decouples the learning procedure of the panoramic reconstruction stage from the translation stage. We show distinct improvements over existing I2I models in translating the StreetLearn dataset in the daytime into diverse conditions. The code will be publicly available online for our community.

Sungdong Kim, Jin-Hwa Kim, Jiyoung Lee et al.

Efficient video-language modeling should consider the computational cost because of a large, sometimes intractable, number of video frames. Parametric approaches such as the attention mechanism may not be ideal since its computational cost quadratically increases as the video length increases. Rather, previous studies have relied on offline feature extraction or frame sampling to represent the video efficiently, focusing on cross-modal modeling in short video clips. In this paper, we propose a semi-parametric video-grounded text generation model, SeViT, a novel perspective on scalable video-language modeling toward long untrimmed videos. Treating a video as an external data store, SeViT includes a non-parametric frame retriever to select a few query-relevant frames from the data store for a given query and a parametric generator to effectively aggregate the frames with the query via late fusion methods. Experimental results demonstrate our method has a significant advantage in longer videos and causal video understanding. Moreover, our model achieves the new state of the art on four video-language datasets, iVQA (+4.8), Next-QA (+6.9), and Activitynet-QA (+4.8) in accuracy, and MSRVTT-Caption (+3.6) in CIDEr.

Jin-Hwa Kim, Yunji Kim, Jiyoung Lee et al.

Text-to-image generation and image captioning are recently emerged as a new experimental paradigm to assess machine intelligence. They predict continuous quantity accompanied by their sampling techniques in the generation, making evaluation complicated and intractable to get marginal distributions. Based on a recent trend that multimodal generative evaluations exploit a vison-and-language pre-trained model, we propose the negative Gaussian cross-mutual information using the CLIP features as a unified metric, coined by Mutual Information Divergence (MID). To validate, we extensively compare it with competing metrics using carefully-generated or human-annotated judgments in text-to-image generation and image captioning tasks. The proposed MID significantly outperforms the competitive methods by having consistency across benchmarks, sample parsimony, and robustness toward the exploited CLIP model. We look forward to seeing the underrepresented implications of the Gaussian cross-mutual information in multimodal representation learning and the future works based on this novel proposition.

Kwonyoung Kim, Jungin Park, Jiyoung Lee et al.

Online stereo adaptation tackles the domain shift problem, caused by different environments between synthetic (training) and real (test) datasets, to promptly adapt stereo models in dynamic real-world applications such as autonomous driving. However, previous methods often fail to counteract particular regions related to dynamic objects with more severe environmental changes. To mitigate this issue, we propose to incorporate an auxiliary point-selective network into a meta-learning framework, called PointFix, to provide a robust initialization of stereo models for online stereo adaptation. In a nutshell, our auxiliary network learns to fix local variants intensively by effectively back-propagating local information through the meta-gradient for the robust initialization of the baseline model. This network is model-agnostic, so can be used in any kind of architectures in a plug-and-play manner. We conduct extensive experiments to verify the effectiveness of our method under three adaptation settings such as short-, mid-, and long-term sequences. Experimental results show that the proper initialization of the base stereo model by the auxiliary network enables our learning paradigm to achieve state-of-the-art performance at inference.

Hwan Heo, Taekyung Kim, Jiyoung Lee et al.

Multi-resolution hash encoding has recently been proposed to reduce the computational cost of neural renderings, such as NeRF. This method requires accurate camera poses for the neural renderings of given scenes. However, contrary to previous methods jointly optimizing camera poses and 3D scenes, the naive gradient-based camera pose refinement method using multi-resolution hash encoding severely deteriorates performance. We propose a joint optimization algorithm to calibrate the camera pose and learn a geometric representation using efficient multi-resolution hash encoding. Showing that the oscillating gradient flows of hash encoding interfere with the registration of camera poses, our method addresses the issue by utilizing smooth interpolation weighting to stabilize the gradient oscillation for the ray samplings across hash grids. Moreover, the curriculum training procedure helps to learn the level-wise hash encoding, further increasing the pose refinement. Experiments on the novel-view synthesis datasets validate that our learning frameworks achieve state-of-the-art performance and rapid convergence of neural rendering, even when initial camera poses are unknown.

Jiyoung Lee, Hantae Kim, Hyunchang Cho et al.

Multi-domain Neural Machine Translation (NMT) trains a single model with multiple domains. It is appealing because of its efficacy in handling multiple domains within one model. An ideal multi-domain NMT should learn distinctive domain characteristics simultaneously, however, grasping the domain peculiarity is a non-trivial task. In this paper, we investigate domain-specific information through the lens of mutual information (MI) and propose a new objective that penalizes low MI to become higher. Our method achieved the state-of-the-art performance among the current competitive multi-domain NMT models. Also, we empirically show our objective promotes low MI to be higher resulting in domain-specialized multi-domain NMT.

Radhika Dua, Jiyoung Lee, Joon-myoung Kwon et al.

Electrocardiogram (ECG) signals are beneficial in diagnosing cardiovascular diseases, which are one of the leading causes of death. However, they are often contaminated by noise artifacts and affect the automatic and manual diagnosis process. Automatic deep learning-based examination of ECG signals can lead to inaccurate diagnosis, and manual analysis involves rejection of noisy ECG samples by clinicians, which might cost extra time. To address this limitation, we present a two-stage deep learning-based framework to automatically detect the noisy ECG samples. Through extensive experiments and analysis on two different datasets, we observe that the deep learning-based framework can detect slightly and highly noisy ECG samples effectively. We also study the transfer of the model learned on one dataset to another dataset and observe that the framework effectively detects noisy ECG samples.

Yujin Jeong, Yunji Kim, Sanghyuk Chun et al.

Despite the impressive progress of multimodal generative models, video-to-audio generation still suffers from limited performance and limits the flexibility to prioritize sound synthesis for specific objects within the scene. Conversely, text-to-audio generation methods generate high-quality audio but pose challenges in ensuring comprehensive scene depiction and time-varying control. To tackle these challenges, we propose a novel video-and-text-to-audio generation method, called \ours, where video serves as a conditional control for a text-to-audio generation model. Especially, our method estimates the structural information of sound (namely, energy) from the video while receiving key content cues from a user prompt. We employ a well-performing text-to-audio model to consolidate the video control, which is much more efficient for training multimodal diffusion models with massive triplet-paired (audio-video-text) data. In addition, by separating the generative components of audio, it becomes a more flexible system that allows users to freely adjust the energy, surrounding environment, and primary sound source according to their preferences. Experimental results demonstrate that our method shows superiority in terms of quality, controllability, and training efficiency. Code and demo are available at https://naver-ai.github.io/rewas.

Jiyoung Lee, Minwoo Kim, Seungho Kim et al.

For Large Language Models (LLMs) to be effectively deployed in a specific country, they must possess an understanding of the nation's culture and basic knowledge. To this end, we introduce National Alignment, which measures an alignment between an LLM and a targeted country from two aspects: social value alignment and common knowledge alignment. Social value alignment evaluates how well the model understands nation-specific social values, while common knowledge alignment examines how well the model captures basic knowledge related to the nation. We constructed KorNAT, the first benchmark that measures national alignment with South Korea. For the social value dataset, we obtained ground truth labels from a large-scale survey involving 6,174 unique Korean participants. For the common knowledge dataset, we constructed samples based on Korean textbooks and GED reference materials. KorNAT contains 4K and 6K multiple-choice questions for social value and common knowledge, respectively. Our dataset creation process is meticulously designed and based on statistical sampling theory and was refined through multiple rounds of human review. The experiment results of seven LLMs reveal that only a few models met our reference score, indicating a potential for further enhancement. KorNAT has received government approval after passing an assessment conducted by a government-affiliated organization dedicated to evaluating dataset quality. Samples and detailed evaluation protocols of our dataset can be found in https://huggingface.co/datasets/jiyounglee0523/KorNAT .

Hyemin Boo, Eunsang Lee, Jiyoung Lee

Since deepfakes generated by advanced generative models have rapidly posed serious threats, existing audiovisual deepfake detection approaches struggle to generalize to unseen forgeries. We propose a novel reference-aware audiovisual deepfake detection method, called Referee. Speaker-specific cues from only one-shot examples are leveraged to detect manipulations beyond spatiotemporal artifacts. By matching and aligning identity-related queries from reference and target content into cross-modal features, Referee jointly reasons about audiovisual synchrony and identity consistency. Extensive experiments on FakeAVCeleb, FaceForensics++, and KoDF demonstrate that Referee achieves state-of-the-art performance on cross-dataset and cross-language evaluation protocols. Experimental results highlight the importance of cross-modal identity verification for future deepfake detection. The code is available at https://github.com/ewha-mmai/referee.

Eunbi Choi, Kibong Choi, Seokhee Hong et al.

This technical report presents K-EXAONE, a large-scale multilingual language model developed by LG AI Research. K-EXAONE is built on a Mixture-of-Experts architecture with 236B total parameters, activating 23B parameters during inference. It supports a 256K-token context window and covers six languages: Korean, English, Spanish, German, Japanese, and Vietnamese. We evaluate K-EXAONE on a comprehensive benchmark suite spanning reasoning, agentic, general, Korean, and multilingual abilities. Across these evaluations, K-EXAONE demonstrates performance comparable to open-weight models of similar size. K-EXAONE, designed to advance AI for a better life, is positioned as a powerful proprietary AI foundation model for a wide range of industrial and research applications.

Eunbi Choi, Kibong Choi, Sehyun Chun et al.

This technical report introduces EXAONE 4.5, the first open-weight vision language model released by LG AI Research. EXAONE 4.5 is architected by integrating a dedicated visual encoder into the existing EXAONE 4.0 framework, enabling native multimodal pretraining over both visual and textual modalities. The model is trained on large-scale data with careful curation, particularly emphasizing document-centric corpora that align with LG's strategic application domains. This targeted data design enables substantial performance gains in document understanding and related tasks, while also delivering broad improvements across general language capabilities. EXAONE 4.5 extends context length up to 256K tokens, facilitating long-context reasoning and enterprise-scale use cases. Comparative evaluations demonstrate that EXAONE 4.5 achieves competitive performance in general benchmarks while outperforming state-of-the-art models of similar scale in document understanding and Korean contextual reasoning. As part of LG's ongoing effort toward practical industrial deployment, EXAONE 4.5 is designed to be continuously extended with additional domains and application scenarios to advance AI for a better life.

Jiyoung Lee, Seungho Kim, Jieun Han et al.

Large Language Models (LLMs) are predominantly evaluated on Standard American English (SAE), often overlooking the diversity of global English varieties. This narrow focus may raise fairness concerns as degraded performance on non-standard varieties can lead to unequal benefits for users worldwide. Therefore, it is critical to extensively evaluate the linguistic robustness of LLMs on multiple non-standard English varieties. We introduce Trans-EnV, a framework that automatically transforms SAE datasets into multiple English varieties to evaluate the linguistic robustness. Our framework combines (1) linguistics expert knowledge to curate variety-specific features and transformation guidelines from linguistic literature and corpora, and (2) LLM-based transformations to ensure both linguistic validity and scalability. Using Trans-EnV, we transform six benchmark datasets into 38 English varieties and evaluate seven state-of-the-art LLMs. Our results reveal significant performance disparities, with accuracy decreasing by up to 46.3% on non-standard varieties. These findings highlight the importance of comprehensive linguistic robustness evaluation across diverse English varieties. Each construction of Trans-EnV was validated through rigorous statistical testing and consultation with a researcher in the field of second language acquisition, ensuring its linguistic validity. Our code and datasets are publicly available at https://github.com/jiyounglee-0523/TransEnV and https://huggingface.co/collections/jiyounglee0523/transenv-681eadb3c0c8cf363b363fb1.

Jungin Park, Jiyoung Lee, Kwanghoon Sohn

View-invariant representation learning from egocentric (first-person, ego) and exocentric (third-person, exo) videos is a promising approach toward generalizing video understanding systems across multiple viewpoints. However, this area has been underexplored due to the substantial differences in perspective, motion patterns, and context between ego and exo views. In this paper, we propose a novel masked ego-exo modeling that promotes both causal temporal dynamics and cross-view alignment, called Bootstrap Your Own Views (BYOV), for fine-grained view-invariant video representation learning from unpaired ego-exo videos. We highlight the importance of capturing the compositional nature of human actions as a basis for robust cross-view understanding. Specifically, self-view masking and cross-view masking predictions are designed to learn view-invariant and powerful representations concurrently. Experimental results demonstrate that our BYOV significantly surpasses existing approaches with notable gains across all metrics in four downstream ego-exo video tasks. The code is available at https://github.com/park-jungin/byov.

Junwon Lee, Juhan Nam, Jiyoung Lee

This work introduces a new task, text-conditioned selective video-to-audio (V2A) generation, which produces only the user-intended sound from a multi-object video. This capability is especially crucial in multimedia production, where audio tracks are handled individually for each sound source for precise editing, mixing, and creative control. However, current approaches generate single source-mixed sounds at once, largely because visual features are entangled, and region cues or prompts often fail to specify the source. We propose SelVA, a novel text-conditioned V2A model that treats the text prompt as an explicit selector of target source and modulates video encoder to distinctly extract prompt-relevant video features. The proposed supplementary tokens promote cross-attention by suppressing text-irrelevant activations with efficient parameter tuning, yielding robust semantic and temporal grounding. SelVA further employs a self-augmentation scheme to overcome the lack of mono audio track supervision. We evaluate SelVA on VGG-MONOAUDIO, a curated benchmark of clean single-source videos for such a task. Extensive experiments and ablations consistently verify its effectiveness across audio quality, semantic alignment, and temporal synchronization. Code and demo are available at https://jnwnlee.github.io/selva-demo/.

Jungin Park, Jiyoung Lee, Kwanghoon Sohn

Video summarization aims to select keyframes that are visually diverse and can represent the whole story of a given video. Previous approaches have focused on global interlinkability between frames in a video by temporal modeling. However, fine-grained visual entities, such as objects, are also highly related to the main content of the video. Moreover, language-guided video summarization, which has recently been studied, requires a comprehensive linguistic understanding of complex real-world videos. To consider how all the objects are semantically related to each other, this paper regards video summarization as a language-guided spatiotemporal graph modeling problem. We present recursive spatiotemporal graph networks, called VideoGraph, which formulate the objects and frames as nodes of the spatial and temporal graphs, respectively. The nodes in each graph are connected and aggregated with graph edges, representing the semantic relationships between the nodes. To prevent the edges from being configured with visual similarity, we incorporate language queries derived from the video into the graph node representations, enabling them to contain semantic knowledge. In addition, we adopt a recursive strategy to refine initial graphs and correctly classify each frame node as a keyframe. In our experiments, VideoGraph achieves state-of-the-art performance on several benchmarks for generic and query-focused video summarization in both supervised and unsupervised manners. The code is available at https://github.com/park-jungin/videograph.

Soyoung Yang, Hojun Cho, Jiyoung Lee et al.

Aspect-based sentiment analysis (ABSA) is a challenging task of extracting sentiments along with their corresponding aspects and opinion terms from the text. The inherent subjectivity of span annotation makes variability in the surface forms of extracted terms, complicating the evaluation process. Traditional evaluation methods often constrain ground truths (GT) to a single term, potentially misrepresenting the accuracy of semantically valid predictions that differ in surface form. To address this limitation, we propose a novel and fully automated pipeline that expands existing evaluation sets by adding alternative valid terms for aspect and opinion. Our approach facilitates an equitable assessment of language models by accommodating multiple-answer candidates, resulting in enhanced human agreement compared to single-answer test sets (achieving up to a 10\%p improvement in Kendall's Tau score). Experimental results demonstrate that our expanded evaluation set helps uncover the capabilities of large language models (LLMs) in ABSA tasks, which is concealed by the single-answer GT sets. Consequently, our work contributes to the development of a flexible evaluation framework for ABSA by embracing diverse surface forms to span extraction tasks in a cost-effective and reproducible manner. Our code and dataset is open at https://github.com/dudrrm/zoom-in-n-out-absa.

Jiyoung Lee, Song Park, Sanghyuk Chun et al.

This paper proposes VoxStudio, the first unified and end-to-end speech-to-image model that generates expressive images directly from spoken descriptions by jointly aligning linguistic and paralinguistic information. At its core is a speech information bottleneck (SIB) module, which compresses raw speech into compact semantic tokens, preserving prosody and emotional nuance. By operating directly on these tokens, VoxStudio eliminates the need for an additional speech-to-text system, which often ignores the hidden details beyond text, e.g., tone or emotion. We also release VoxEmoset, a large-scale paired emotional speech-image dataset built via an advanced TTS engine to affordably generate richly expressive utterances. Comprehensive experiments on the SpokenCOCO, Flickr8kAudio, and VoxEmoset benchmarks demonstrate the feasibility of our method and highlight key challenges, including emotional consistency and linguistic ambiguity, paving the way for future research.

Jungin Park, Jiyoung Lee, Kwanghoon Sohn

This paper introduces VLAP, a novel approach that bridges pretrained vision models and large language models (LLMs) to make frozen LLMs understand the visual world. VLAP transforms the embedding space of pretrained vision models into the LLMs' word embedding space using a single linear layer for efficient and general-purpose visual and language understanding. Specifically, we harness well-established word embeddings to bridge two modality embedding spaces. The visual and text representations are simultaneously assigned to a set of word embeddings within pretrained LLMs by formulating the assigning procedure as an optimal transport problem. We predict the assignment of one modality from the representation of another modality data, enforcing consistent assignments for paired multimodal data. This allows vision and language representations to contain the same information, grounding the frozen LLMs' word embedding space in visual data. Moreover, a robust semantic taxonomy of LLMs can be preserved with visual data since the LLMs interpret and reason linguistic information from correlations between word embeddings. Experimental results show that VLAP achieves substantial improvements over the previous linear transformation-based approaches across a range of vision-language tasks, including image captioning, visual question answering, and cross-modal retrieval. We also demonstrate the learned visual representations hold a semantic taxonomy of LLMs, making visual semantic arithmetic possible.

Hyemin Boo, Hyoryung Kim, Myungjin Lee et al.

Text-to-image diffusion models generate realistic and coherent images but often fail to follow numerical instructions in text, revealing a gap between language and visual representation. Interestingly, we found that these models are not entirely blind to numbers-they are implicitly aware of their own counting accuracy, as their internal signals shift in consistent ways depending on whether the output meets the specified count. This observation suggests that the model already encodes a latent notion of numerical correctness, which can be harnessed to guide generation more precisely. Building on this intuition, we introduce CountSteer, a training-free method that improves generation of specified object counts by steering the model's cross-attention hidden states during inference. In our experiments, CountSteer improved object-count accuracy by about 4% without compromising visual quality, demonstrating a simple yet effective step toward more controllable and semantically reliable text-to-image generation.

Jinbae Seo, Hyeongjun Kwon, Kwonyoung Kim et al.

Audiovisual instance segmentation (AVIS) requires accurately localizing and tracking sounding objects throughout video sequences. Existing methods suffer from visual bias stemming from two fundamental issues: uniform additive fusion prevents queries from specializing to different sound sources, while visual-only training objectives allow queries to converge to arbitrary salient objects. We propose Audio-Centric Query Generation using cross-attention, enabling each query to selectively attend to distinct sound sources and carry sound-specific priors into visual decoding. Additionally, we introduce Sound-Aware Ordinal Counting (SAOC) loss that explicitly supervises sounding object numbers through ordinal regression with monotonic consistency constraints, preventing visual-only convergence during training. Experiments on AVISeg benchmark demonstrate consistent improvements: +1.64 mAP, +0.6 HOTA, and +2.06 FSLA, validating that query specialization and explicit counting supervision are crucial for accurate audiovisual instance segmentation.

Jinhyun Jang, Jiyoung Lee, Kwanghoon Sohn

Image-text matching aims to build correspondences between visual and textual data by learning their pairwise similarities. Most existing approaches have adopted sparse binary supervision, indicating whether a pair of images and sentences matches or not. However, such sparse supervision covers a limited subset of image-text relationships, neglecting their inherent many-to-many correspondences; an image can be described in numerous texts at different descriptive levels. Moreover, existing approaches overlook the implicit connections from general to specific descriptions, which form the underlying rationale for the many-to-many relationships between vision and language. In this work, we propose descriptive image-text matching, called DITM, to learn the graded contextual similarity between image and text by exploring the descriptive flexibility of language. We formulate the descriptiveness score of each sentence with cumulative term frequency-inverse document frequency (TF-IDF) to balance the pairwise similarity according to the keywords in the sentence. Our method leverages sentence descriptiveness to learn robust image-text matching in two key ways: (1) to refine the false negative labeling, dynamically relaxing the connectivity between positive and negative pairs, and (2) to build more precise matching, aligning a set of relevant sentences in a generic-to-specific order. By moving beyond rigid binary supervision, DITM enhances the discovery of both optimal matches and potential positive pairs. Extensive experiments on MS-COCO, Flickr30K, and CxC datasets demonstrate the effectiveness of our method in representing complex image-text relationships compared to state-of-the-art approaches. In addition, DITM enhances the hierarchical reasoning ability of the model, supported by the extensive analysis on HierarCaps benchmark.

Somi Jeong, Jiyoung Lee, Kwanghoon Sohn

Over the past few years, image-to-image (I2I) translation methods have been proposed to translate a given image into diverse outputs. Despite the impressive results, they mainly focus on the I2I translation between two domains, so the multi-domain I2I translation still remains a challenge. To address this problem, we propose a novel multi-domain unsupervised image-to-image translation (MDUIT) framework that leverages the decomposed content feature and appearance adaptive convolution to translate an image into a target appearance while preserving the given geometric content. We also exploit a contrast learning objective, which improves the disentanglement ability and effectively utilizes multi-domain image data in the training process by pairing the semantically similar images. This allows our method to learn the diverse mappings between multiple visual domains with only a single framework. We show that the proposed method produces visually diverse and plausible results in multiple domains compared to the state-of-the-art methods.

Woncheol Shin, Gyubok Lee, Jiyoung Lee et al.

This is an exploratory study that discovers the current image quantization (vector quantization) do not satisfy translation equivariance in the quantized space due to aliasing. Instead of focusing on anti-aliasing, we propose a simple yet effective way to achieve translation-equivariant image quantization by enforcing orthogonality among the codebook embeddings. To explore the advantages of translation-equivariant image quantization, we conduct three proof-of-concept experiments with a carefully controlled dataset: (1) text-to-image generation, where the quantized image indices are the target to predict, (2) image-to-text generation, where the quantized image indices are given as a condition, (3) using a smaller training set to analyze sample efficiency. From the strictly controlled experiments, we empirically verify that the translation-equivariant image quantizer improves not only sample efficiency but also the accuracy over VQGAN up to +11.9% in text-to-image generation and +3.9% in image-to-text generation.

Kyunghoon Hur, Jiyoung Lee, Jungwoo Oh et al.

EHR systems lack a unified code system forrepresenting medical concepts, which acts asa barrier for the deployment of deep learningmodels in large scale to multiple clinics and hos-pitals. To overcome this problem, we introduceDescription-based Embedding,DescEmb, a code-agnostic representation learning framework forEHR. DescEmb takes advantage of the flexibil-ity of neural language understanding models toembed clinical events using their textual descrip-tions rather than directly mapping each event toa dedicated embedding. DescEmb outperformedtraditional code-based embedding in extensiveexperiments, especially in a zero-shot transfertask (one hospital to another), and was able totrain a single unified model for heterogeneousEHR datasets.

Jiyoung Lee, Wonjae Kim, Daehoon Gwak et al.

Periodic signals play an important role in daily lives. Although conventional sequential models have shown remarkable success in various fields, they still come short in modeling periodicity; they either collapse, diverge or ignore details. In this paper, we introduce a novel framework inspired by Fourier series to generate periodic signals. We first decompose the given signals into multiple sines and cosines and then conditionally generate periodic signals with the output components. We have shown our model efficacy on three tasks: reconstruction, imputation and conditional generation. Our model outperforms baselines in all tasks and shows more stable and refined results.

Jaehoon Cho, Jiyoung Lee, Changjae Oh et al.

Video prediction, forecasting the future frames from a sequence of input frames, is a challenging task since the view changes are influenced by various factors, such as the global context surrounding the scene and local motion dynamics. In this paper, we propose a new framework to integrate these complementary attributes to predict complex pixel dynamics through deep networks. We present global context propagation networks that iteratively aggregate the non-local neighboring representations to preserve the contextual information over the past frames. To capture the local motion pattern of objects, we also devise local filter memory networks that generate adaptive filter kernels by storing the prototypical motion of moving objects in the memory. The proposed framework, utilizing the outputs from both networks, can address blurry predictions and color distortion. We conduct experiments on Caltech pedestrian and UCF101 datasets, and demonstrate state-of-the-art results. Especially for multi-step prediction, we obtain an outstanding performance in quantitative and qualitative evaluation.

Jin Kim, Jiyoung Lee, Jungin Park et al.

Domain generalization aims to learn a prediction model on multi-domain source data such that the model can generalize to a target domain with unknown statistics. Most existing approaches have been developed under the assumption that the source data is well-balanced in terms of both domain and class. However, real-world training data collected with different composition biases often exhibits severe distribution gaps for domain and class, leading to substantial performance degradation. In this paper, we propose a self-balanced domain generalization framework that adaptively learns the weights of losses to alleviate the bias caused by different distributions of the multi-domain source data. The self-balanced scheme is based on an auxiliary reweighting network that iteratively updates the weight of loss conditioned on the domain and class information by leveraging balanced meta data. Experimental results demonstrate the effectiveness of our method overwhelming state-of-the-art works for domain generalization.

Kyunghoon Hur, Jiyoung Lee, Jungwoo Oh et al.

Substantial increase in the use of Electronic Health Records (EHRs) has opened new frontiers for predictive healthcare. However, while EHR systems are nearly ubiquitous, they lack a unified code system for representing medical concepts. Heterogeneous formats of EHR present a substantial barrier for the training and deployment of state-of-the-art deep learning models at scale. To overcome this problem, we introduce Description-based Embedding, DescEmb, a code-agnostic description-based representation learning framework for predictive modeling on EHR. DescEmb takes advantage of the flexibility of neural language understanding models while maintaining a neutral approach that can be combined with prior frameworks for task-specific representation learning or predictive modeling. We tested our model's capacity on various experiments including prediction tasks, transfer learning and pooled learning. DescEmb shows higher performance in overall experiments compared to code-based approach, opening the door to a text-based approach in predictive healthcare research that is not constrained by EHR structure nor special domain knowledge.

Daniel McDuff, Yale Song, Jiyoung Lee et al.

The ability to perform causal and counterfactual reasoning are central properties of human intelligence. Decision-making systems that can perform these types of reasoning have the potential to be more generalizable and interpretable. Simulations have helped advance the state-of-the-art in this domain, by providing the ability to systematically vary parameters (e.g., confounders) and generate examples of the outcomes in the case of counterfactual scenarios. However, simulating complex temporal causal events in multi-agent scenarios, such as those that exist in driving and vehicle navigation, is challenging. To help address this, we present a high-fidelity simulation environment that is designed for developing algorithms for causal discovery and counterfactual reasoning in the safety-critical context. A core component of our work is to introduce \textit{agency}, such that it is simple to define and create complex scenarios using high-level definitions. The vehicles then operate with agency to complete these objectives, meaning low-level behaviors need only be controlled if necessary. We perform experiments with three state-of-the-art methods to create baselines and highlight the affordances of this environment. Finally, we highlight challenges and opportunities for future work.

Jungin Park, Jiyoung Lee, Kwanghoon Sohn

This paper presents a novel method, termed Bridge to Answer, to infer correct answers for questions about a given video by leveraging adequate graph interactions of heterogeneous crossmodal graphs. To realize this, we learn question conditioned visual graphs by exploiting the relation between video and question to enable each visual node using question-to-visual interactions to encompass both visual and linguistic cues. In addition, we propose bridged visual-to-visual interactions to incorporate two complementary visual information on appearance and motion by placing the question graph as an intermediate bridge. This bridged architecture allows reliable message passing through compositional semantics of the question to generate an appropriate answer. As a result, our method can learn the question conditioned visual representations attributed to appearance and motion that show powerful capability for video question answering. Extensive experiments prove that the proposed method provides effective and superior performance than state-of-the-art methods on several benchmarks.

Jiyoung Lee, Soo-Whan Chung, Sunok Kim et al.

In this paper, we address the problem of separating individual speech signals from videos using audio-visual neural processing. Most conventional approaches utilize frame-wise matching criteria to extract shared information between co-occurring audio and video. Thus, their performance heavily depends on the accuracy of audio-visual synchronization and the effectiveness of their representations. To overcome the frame discontinuity problem between two modalities due to transmission delay mismatch or jitter, we propose a cross-modal affinity network (CaffNet) that learns global correspondence as well as locally-varying affinities between audio and visual streams. Given that the global term provides stability over a temporal sequence at the utterance-level, this resolves the label permutation problem characterized by inconsistent assignments. By extending the proposed cross-modal affinity on the complex network, we further improve the separation performance in the complex spectral domain. Experimental results verify that the proposed methods outperform conventional ones on various datasets, demonstrating their advantages in real-world scenarios.

Jungin Park, Jiyoung Lee, Ig-Jae Kim et al.

The goal of video summarization is to select keyframes that are visually diverse and can represent a whole story of an input video. State-of-the-art approaches for video summarization have mostly regarded the task as a frame-wise keyframe selection problem by aggregating all frames with equal weight. However, to find informative parts of the video, it is necessary to consider how all the frames of the video are related to each other. To this end, we cast video summarization as a graph modeling problem. We propose recursive graph modeling networks for video summarization, termed SumGraph, to represent a relation graph, where frames are regarded as nodes and nodes are connected by semantic relationships among frames. Our networks accomplish this through a recursive approach to refine an initially estimated graph to correctly classify each node as a keyframe by reasoning the graph representation via graph convolutional networks. To leverage SumGraph in a more practical environment, we also present a way to adapt our graph modeling in an unsupervised fashion. With SumGraph, we achieved state-of-the-art performance on several benchmarks for video summarization in both supervised and unsupervised manners.

Jiyoung Lee, Seungryong Kim, Sunok Kim et al.

Traditional techniques for emotion recognition have focused on the facial expression analysis only, thus providing limited ability to encode context that comprehensively represents the emotional responses. We present deep networks for context-aware emotion recognition, called CAER-Net, that exploit not only human facial expression but also context information in a joint and boosting manner. The key idea is to hide human faces in a visual scene and seek other contexts based on an attention mechanism. Our networks consist of two sub-networks, including two-stream encoding networks to seperately extract the features of face and context regions, and adaptive fusion networks to fuse such features in an adaptive fashion. We also introduce a novel benchmark for context-aware emotion recognition, called CAER, that is more appropriate than existing benchmarks both qualitatively and quantitatively. On several benchmarks, CAER-Net proves the effect of context for emotion recognition. Our dataset is available at http://caer-dataset.github.io.