Heung-Seon Oh

Junuk Jung, Seonhoon Lee, Heung-Seon Oh et al.

The goal of face recognition (FR) can be viewed as a pair similarity optimization problem, maximizing a similarity set $\mathcal{S}^p$ over positive pairs, while minimizing similarity set $\mathcal{S}^n$ over negative pairs. Ideally, it is expected that FR models form a well-discriminative feature space (WDFS) that satisfies $\inf{\mathcal{S}^p} > \sup{\mathcal{S}^n}$. With regard to WDFS, the existing deep feature learning paradigms (i.e., metric and classification losses) can be expressed as a unified perspective on different pair generation (PG) strategies. Unfortunately, in the metric loss (ML), it is infeasible to generate negative pairs taking all classes into account in each iteration because of the limited mini-batch size. In contrast, in classification loss (CL), it is difficult to generate extremely hard negative pairs owing to the convergence of the class weight vectors to their center. This leads to a mismatch between the two similarity distributions of the sampled pairs and all negative pairs. Thus, this paper proposes a unified negative pair generation (UNPG) by combining two PG strategies (i.e., MLPG and CLPG) from a unified perspective to alleviate the mismatch. UNPG introduces useful information about negative pairs using MLPG to overcome the CLPG deficiency. Moreover, it includes filtering the similarities of noisy negative pairs to guarantee reliable convergence and improved performance. Exhaustive experiments show the superiority of UNPG by achieving state-of-the-art performance across recent loss functions on public benchmark datasets. Our code and pretrained models are publicly available.

Euihyeok Lee, Seonghyeon Kim, SangHun Im et al.

Self-talk-an internal dialogue that can occur silently or be spoken aloud-plays a crucial role in emotional regulation, cognitive processing, and motivation, yet has remained largely invisible and unmeasurable in everyday life. In this paper, we present MutterMeter, a mobile system that automatically detects vocalized self-talk from audio captured by earable microphones in real-world settings. Detecting self-talk is technically challenging due to its diverse acoustic forms, semantic and grammatical incompleteness, and irregular occurrence patterns, which differ fundamentally from assumptions underlying conventional speech understanding models. To address these challenges, MutterMeter employs a hierarchical classification architecture that progressively integrates acoustic, linguistic, and contextual information through a sequential processing pipeline, adaptively balancing accuracy and computational efficiency. We build and evaluate MutterMeter using a first-of-its-kind dataset comprising 31.1 hours of audio collected from 25 participants. Experimental results demonstrate that MutterMeter achieves robust performance with a macro-averaged F1 score of 0.84, outperforming conventional approaches, including LLM-based and speech emotion recognition models.

Kyungmin Kim, SangHun Im, GiBaeg Kim et al.

Text augmentation (TA) is a critical technique for text classification, especially in few-shot settings. This paper introduces a novel LLM-based TA method, TARDiS, to address challenges inherent in the generation and alignment stages of two-stage TA methods. For the generation stage, we propose two generation processes, SEG and CEG, incorporating multiple class-specific prompts to enhance diversity and separability. For the alignment stage, we introduce a class adaptation (CA) method to ensure that generated examples align with their target classes through verification and modification. Experimental results demonstrate TARDiS's effectiveness, outperforming state-of-the-art LLM-based TA methods in various few-shot text classification tasks. An in-depth analysis confirms the detailed behaviors at each stage.

Yeong-Seung Baek, Heung-Seon Oh

3D visual grounding (VG) aims to locate objects or regions within 3D scenes guided by natural language descriptions. While indoor 3D VG has advanced, outdoor 3D VG remains underexplored due to two challenges: (1) large-scale outdoor LiDAR scenes are dominated by background points and contain limited foreground information, making cross-modal alignment and contextual understanding more difficult; and (2) most outdoor datasets lack spatial annotations for referential non-target objects, which hinders explicit learning of referential context. To this end, we propose LidaRefer, a context-aware 3D VG framework for outdoor scenes. LidaRefer incorporates an object-centric feature selection strategy to focus on semantically relevant visual features while reducing computational overhead. Then, its transformer-based encoder-decoder architecture excels at establishing fine-grained cross-modal alignment between refined visual features and word-level text features, and capturing comprehensive global context. Additionally, we present Discriminative-Supportive Collaborative localization (DiSCo), a novel supervision strategy that explicitly models spatial relationships between target, contextual, and ambiguous objects for accurate target identification. To enable this without manual labeling, we introduce a pseudo-labeling approach that retrieves 3D localization labels for referential non-target objects. LidaRefer achieves state-of-the-art performance on Talk2Car-3D dataset under various evaluation settings.

Jongyun Choi, Hyesoo Kong, Hwamook Yoon et al.

The volume of academic literature, such as academic conference papers and journals, has increased rapidly worldwide, and research on metadata extraction is ongoing. However, high-performing metadata extraction is still challenging due to diverse layout formats according to journal publishers. To accommodate the diversity of the layouts of academic journals, we propose a novel LAyout-aware Metadata Extraction (LAME) framework equipped with the three characteristics (e.g., design of an automatic layout analysis, construction of a large meta-data training set, and construction of Layout-MetaBERT). We designed an automatic layout analysis using PDFMiner. Based on the layout analysis, a large volume of metadata-separated training data, including the title, abstract, author name, author affiliated organization, and keywords, were automatically extracted. Moreover, we constructed Layout-MetaBERT to extract the metadata from academic journals with varying layout formats. The experimental results with Layout-MetaBERT exhibited robust performance (Macro-F1, 93.27%) in metadata extraction for unseen journals with different layout formats.

SangHun Im, Gibaeg Kim, Heung-Seon Oh et al.

Hierarchical text classification (HTC) is essential for various real applications. However, HTC models are challenging to develop because they often require processing a large volume of documents and labels with hierarchical taxonomy. Recent HTC models based on deep learning have attempted to incorporate hierarchy information into a model structure. Consequently, these models are challenging to implement when the model parameters increase for a large-scale hierarchy because the model structure depends on the hierarchy size. To solve this problem, we formulate HTC as a sub-hierarchy sequence generation to incorporate hierarchy information into a target label sequence instead of the model structure. Subsequently, we propose the Hierarchy DECoder (HiDEC), which decodes a text sequence into a sub-hierarchy sequence using recursive hierarchy decoding, classifying all parents at the same level into children at once. In addition, HiDEC is trained to use hierarchical path information from a root to each leaf in a sub-hierarchy composed of the labels of a target document via an attention mechanism and hierarchy-aware masking. HiDEC achieved state-of-the-art performance with significantly fewer model parameters than existing models on benchmark datasets, such as RCV1-v2, NYT, and EURLEX57K.

Junuk Jung, Sungbin Son, Joochan Park et al.

The performance of face recognition has become saturated for public benchmark datasets such as LFW, CFP-FP, and AgeDB, owing to the rapid advances in CNNs. However, the effects of faces with various fine-grained conditions on FR models have not been investigated because of the absence of such datasets. This paper analyzes their effects in terms of different conditions and loss functions using K-FACE, a recently introduced FR dataset with fine-grained conditions. We propose a novel loss function, MixFace, that combines classification and metric losses. The superiority of MixFace in terms of effectiveness and robustness is demonstrated experimentally on various benchmark datasets.