arXiv subjects

Yujin Kim

Publications and source records attributed to Yujin Kim.

At least 37 records · Page 2Linked to original sources

arXiv2025

Self-Training Elicits Concise Reasoning in Large Language Models

Chain-of-thought (CoT) reasoning has enabled large language models (LLMs) to utilize additional computation through intermediate tokens to solve complex tasks. However, we posit that typical reasoning traces contain many redundant tokens, incurring extraneous inference costs. Upon examination of the output distribution of current LLMs, we find evidence on their latent ability to reason more concisely, relative to their default behavior. To elicit this capability, we propose simple fine-tuning methods which leverage self-generated concise reasoning paths obtained by best-of-N sampling and few-shot conditioning, in task-specific settings. Our combined method achieves a 30% reduction in output tokens on average, across five model families on GSM8K and MATH, while maintaining average accuracy. By exploiting the fundamental stochasticity and in-context learning capabilities of LLMs, our self-training approach robustly elicits concise reasoning on a wide range of models, including those with extensive post-training. Code is available at https://github.com/TergelMunkhbat/concise-reasoning

cs.CL↗

arXiv2025

Guiding Reasoning in Small Language Models with LLM Assistance

The limited reasoning capabilities of small language models (SLMs) cast doubt on their suitability for tasks demanding deep, multi-step logical deduction. This paper introduces a framework called Small Reasons, Large Hints (SMART), which selectively augments SLM reasoning with targeted guidance from large language models (LLMs). Inspired by the concept of cognitive scaffolding, SMART employs a score-based evaluation to identify uncertain reasoning steps and injects corrective LLM-generated reasoning only when necessary. By framing structured reasoning as an optimal policy search, our approach steers the reasoning trajectory toward correct solutions without exhaustive sampling. Our experiments on mathematical reasoning datasets demonstrate that targeted external scaffolding significantly improves performance, paving the way for collaborative use of both SLM and LLM to tackle complex reasoning tasks that are currently unsolvable by SLMs alone.

cs.CL↗

arXiv2025

Distilling Realizable Students from Unrealizable Teachers

We study policy distillation under privileged information, where a student policy with only partial observations must learn from a teacher with full-state access. A key challenge is information asymmetry: the student cannot directly access the teacher's state space, leading to distributional shifts and policy degradation. Existing approaches either modify the teacher to produce realizable but sub-optimal demonstrations or rely on the student to explore missing information independently, both of which are inefficient. Our key insight is that the student should strategically interact with the teacher --querying only when necessary and resetting from recovery states --to stay on a recoverable path within its own observation space. We introduce two methods: (i) an imitation learning approach that adaptively determines when the student should query the teacher for corrections, and (ii) a reinforcement learning approach that selects where to initialize training for efficient exploration. We validate our methods in both simulated and real-world robotic tasks, demonstrating significant improvements over standard teacher-student baselines in training efficiency and final performance. The project website is available at : https://portal-cornell.github.io/CritiQ_ReTRy/

cs.RO↗

arXiv2025

Evaluating Visual Explanations of Attention Maps for Transformer-based Medical Imaging

Although Vision Transformers (ViTs) have recently demonstrated superior performance in medical imaging problems, they face explainability issues similar to previous architectures such as convolutional neural networks. Recent research efforts suggest that attention maps, which are part of decision-making process of ViTs can potentially address the explainability issue by identifying regions influencing predictions, especially in models pretrained with self-supervised learning. In this work, we compare the visual explanations of attention maps to other commonly used methods for medical imaging problems. To do so, we employ four distinct medical imaging datasets that involve the identification of (1) colonic polyps, (2) breast tumors, (3) esophageal inflammation, and (4) bone fractures and hardware implants. Through large-scale experiments on the aforementioned datasets using various supervised and self-supervised pretrained ViTs, we find that although attention maps show promise under certain conditions and generally surpass GradCAM in explainability, they are outperformed by transformer-specific interpretability methods. Our findings indicate that the efficacy of attention maps as a method of interpretability is context-dependent and may be limited as they do not consistently provide the comprehensive insights required for robust medical decision-making.

cs.CV↗

arXiv2025

I am not thinking anymore, just following the path.: Investigating Task Delegation Trend of Author-AI Co-Creation with Generative AIs

This paper investigates the task delegation trends of digital comic authors to generative AIs during the creation process. We observed 16 digital comic authors using generative AIs during the drafting stage. We categorized authors delegation levels and examined the extent of delegation, variations in AI usage, and calibration of delegation in co-creation. Our findings show that most authors delegate significant tasks to AI, with higher delegation linked to less time spent on creation and more detailed questions to AI. After co-creation, about 60% of authors adjusted their delegation levels, mostly calibrating to less delegation due to loss of agency and AIs unoriginal outputs. We suggest strategies for calibrating delegation to an appropriate level, redefine trust in human-AI co-creation, and propose novel measurements for trust in these contexts. Our study provides insights into how authors can effectively collaborate with generative AIs, balance delegation, and navigate AIs role in the creative process.

cs.HC↗

arXiv2024

Subspace-wise Hybrid RL for Articulated Object Manipulation

Articulated object manipulation is a challenging task, requiring constrained motion and adaptive control to handle the unknown dynamics of the manipulated objects. While reinforcement learning (RL) has been widely employed to tackle various scenarios and types of articulated objects, the complexity of these tasks, stemming from multiple intertwined objectives makes learning a control policy in the full task space highly difficult. To address this issue, we propose a Subspace-wise hybrid RL (SwRL) framework that learns policies for each divided task space, or subspace, based on independent objectives. This approach enables adaptive force modulation to accommodate the unknown dynamics of objects. Additionally, it effectively leverages the previously underlooked redundant subspace, thereby maximizing the robot's dexterity. Our method enhances both learning efficiency and task execution performance, as validated through simulations and real-world experiments. Supplementary video is available at https://youtu.be/PkNxv0P8Atk

cs.RO↗

arXiv2024

Expressivity of deterministic quantum computation with one qubit

Deterministic quantum computation with one qubit (DQC1) is of significant theoretical and practical interest due to its computational advantages in certain problems, despite its subuniversality with limited quantum resources. In this work, we introduce parameterized DQC1 as a quantum machine learning model. We demonstrate that the gradient of the measurement outcome of a DQC1 circuit with respect to its gate parameters can be computed directly using the DQC1 protocol. This allows for gradient-based optimization of DQC1 circuits, positioning DQC1 as the sole quantum protocol for both training and inference. We then analyze the expressivity of the parameterized DQC1 circuits, characterizing the set of learnable functions, and show that DQC1-based machine learning (ML) is as powerful as quantum neural networks based on universal computation. Our findings highlight the potential of DQC1 as a practical and versatile platform for ML, capable of rivaling more complex quantum computing models while utilizing simpler quantum resources.

quant-ph↗

arXiv2024

BAPO: Base-Anchored Preference Optimization for Overcoming Forgetting in Large Language Models Personalization

While learning to align Large Language Models (LLMs) with human preferences has shown remarkable success, aligning these models to meet the diverse user preferences presents further challenges in preserving previous knowledge. This paper examines the impact of personalized preference optimization on LLMs, revealing that the extent of knowledge loss varies significantly with preference heterogeneity. Although previous approaches have utilized the KL constraint between the reference model and the policy model, we observe that they fail to maintain general knowledge and alignment when facing personalized preferences. To this end, we introduce Base-Anchored Preference Optimization (BAPO), a simple yet effective approach that utilizes the initial responses of reference model to mitigate forgetting while accommodating personalized alignment. BAPO effectively adapts to diverse user preferences while minimally affecting global knowledge or general alignment. Our experiments demonstrate the efficacy of BAPO in various setups.

cs.AI↗

arXiv2024

Carpe Diem: On the Evaluation of World Knowledge in Lifelong Language Models

The dynamic nature of knowledge in an ever-changing world presents challenges for language models trained on static data; the model in the real world often requires not only acquiring new knowledge but also overwriting outdated information into updated ones. To study the ability of language models for these time-dependent dynamics in human language, we introduce a novel task, EvolvingQA, a temporally evolving question-answering benchmark designed for training and evaluating LMs on an evolving Wikipedia database. The construction of EvolvingQA is automated with our pipeline using large language models. We uncover that existing continual learning baselines suffer from updating and removing outdated knowledge. Our analysis suggests that models fail to rectify knowledge due to small weight gradients. In addition, we elucidate that language models particularly struggle to reflect the change of numerical or temporal information. Our work aims to model the dynamic nature of real-world information, suggesting faithful evaluations of the evolution-adaptability of language models.

cs.CL↗

arXiv2024

TnT-LLM: Text Mining at Scale with Large Language Models

Transforming unstructured text into structured and meaningful forms, organized by useful category labels, is a fundamental step in text mining for downstream analysis and application. However, most existing methods for producing label taxonomies and building text-based label classifiers still rely heavily on domain expertise and manual curation, making the process expensive and time-consuming. This is particularly challenging when the label space is under-specified and large-scale data annotations are unavailable. In this paper, we address these challenges with Large Language Models (LLMs), whose prompt-based interface facilitates the induction and use of large-scale pseudo labels. We propose TnT-LLM, a two-phase framework that employs LLMs to automate the process of end-to-end label generation and assignment with minimal human effort for any given use-case. In the first phase, we introduce a zero-shot, multi-stage reasoning approach which enables LLMs to produce and refine a label taxonomy iteratively. In the second phase, LLMs are used as data labelers that yield training samples so that lightweight supervised classifiers can be reliably built, deployed, and served at scale. We apply TnT-LLM to the analysis of user intent and conversational domain for Bing Copilot (formerly Bing Chat), an open-domain chat-based search engine. Extensive experiments using both human and automatic evaluation metrics demonstrate that TnT-LLM generates more accurate and relevant label taxonomies when compared against state-of-the-art baselines, and achieves a favorable balance between accuracy and efficiency for classification at scale. We also share our practical experiences and insights on the challenges and opportunities of using LLMs for large-scale text mining in real-world applications.

cs.CL↗

arXiv2024

Leveraging Large Language Models for Hybrid Workplace Decision Support

Large Language Models (LLMs) hold the potential to perform a variety of text processing tasks and provide textual explanations for proposed actions or decisions. In the era of hybrid work, LLMs can provide intelligent decision support for workers who are designing their hybrid work plans. In particular, they can offer suggestions and explanations to workers balancing numerous decision factors, thereby enhancing their work experience. In this paper, we present a decision support model for workspaces in hybrid work environments, leveraging the reasoning skill of LLMs. We first examine LLM's capability of making suitable workspace suggestions. We find that its reasoning extends beyond the guidelines in the prompt and the LLM can manage the trade-off among the available resources in the workspaces. We conduct an extensive user study to understand workers' decision process for workspace choices and evaluate the effectiveness of the system. We observe that a worker's decision could be influenced by the LLM's suggestions and explanations. The participants in our study find the system to be convenient, regardless of whether reasons are provided or not. Our results show that employees can benefit from the LLM-empowered system for their workspace selection in hybrid workplace.

cs.CL↗

arXiv2023

A Literature Review on the Smart Wheelchair Systems

This study offers an in-depth analysis of smart wheelchair (SW) systems, charting their progression from early developments to future innovations. It delves into various Brain-Computer Interface (BCI) systems, including mu rhythm, event-related potential, and steady-state visual evoked potential. The paper addresses challenges in signal categorization, proposing the sparse Bayesian extreme learning machine as an innovative solution. Additionally, it explores the integration of emotional states in BCI systems, the application of alternative control methods such as EMG-based systems, and the deployment of intelligent adaptive interfaces utilizing recurrent quantum neural networks. The study also covers advancements in autonomous navigation, assistance, and mapping, emphasizing their importance in SW systems. The human aspect of SW interaction receives considerable attention, specifically in terms of privacy, physiological factors, and the refinement of control mechanisms. The paper acknowledges the commercial challenges faced, like the limitations of indoor usage and the necessity for user training. For future applications, the research explores the potential of autonomous systems adept at adapting to changing environments and user needs. This exploration includes reinforcement learning and various control methods, such as eye and voice control, to improve adaptability and interaction. The potential integration with smart home technologies, including advanced features such as robotic arms, is also considered, aiming to further enhance user accessibility and independence. Ultimately, this study seeks to provide a thorough overview of SW systems, presenting extensive research to detail their historical evolution, current state, and future prospects.

eess.SY↗

arXiv2023

HARE: Explainable Hate Speech Detection with Step-by-Step Reasoning

With the proliferation of social media, accurate detection of hate speech has become critical to ensure safety online. To combat nuanced forms of hate speech, it is important to identify and thoroughly explain hate speech to help users understand its harmful effects. Recent benchmarks have attempted to tackle this issue by training generative models on free-text annotations of implications in hateful text. However, we find significant reasoning gaps in the existing annotations schemes, which may hinder the supervision of detection models. In this paper, we introduce a hate speech detection framework, HARE, which harnesses the reasoning capabilities of large language models (LLMs) to fill these gaps in explanations of hate speech, thus enabling effective supervision of detection models. Experiments on SBIC and Implicit Hate benchmarks show that our method, using model-generated data, consistently outperforms baselines, using existing free-text human annotations. Analysis demonstrates that our method enhances the explanation quality of trained models and improves generalization to unseen datasets. Our code is available at https://github.com/joonkeekim/hare-hate-speech.git.

cs.CL↗

arXiv2023

NASH: A Simple Unified Framework of Structured Pruning for Accelerating Encoder-Decoder Language Models

Structured pruning methods have proven effective in reducing the model size and accelerating inference speed in various network architectures such as Transformers. Despite the versatility of encoder-decoder models in numerous NLP tasks, the structured pruning methods on such models are relatively less explored compared to encoder-only models. In this study, we investigate the behavior of the structured pruning of the encoder-decoder models in the decoupled pruning perspective of the encoder and decoder component, respectively. Our findings highlight two insights: (1) the number of decoder layers is the dominant factor of inference speed, and (2) low sparsity in the pruned encoder network enhances generation quality. Motivated by these findings, we propose a simple and effective framework, NASH, that narrows the encoder and shortens the decoder networks of encoder-decoder models. Extensive experiments on diverse generation and inference tasks validate the effectiveness of our method in both speedup and output quality.

cs.CL↗

arXiv2023

NaturalInversion: Data-Free Image Synthesis Improving Real-World Consistency

We introduce NaturalInversion, a novel model inversion-based method to synthesize images that agrees well with the original data distribution without using real data. In NaturalInversion, we propose: (1) a Feature Transfer Pyramid which uses enhanced image prior of the original data by combining the multi-scale feature maps extracted from the pre-trained classifier, (2) a one-to-one approach generative model where only one batch of images are synthesized by one generator to bring the non-linearity to optimization and to ease the overall optimizing process, (3) learnable Adaptive Channel Scaling parameters which are end-to-end trained to scale the output image channel to utilize the original image prior further. With our NaturalInversion, we synthesize images from classifiers trained on CIFAR-10/100 and show that our images are more consistent with original data distribution than prior works by visualization and additional analysis. Furthermore, our synthesized images outperform prior works on various applications such as knowledge distillation and pruning, demonstrating the effectiveness of our proposed method.

cs.CV↗

arXiv2022

MonaCoBERT: Monotonic attention based ConvBERT for Knowledge Tracing

Knowledge tracing (KT) is a field of study that predicts the future performance of students based on prior performance datasets collected from educational applications such as intelligent tutoring systems, learning management systems, and online courses. Some previous studies on KT have concentrated only on the interpretability of the model, whereas others have focused on enhancing the performance. Models that consider both interpretability and the performance improvement have been insufficient. Moreover, models that focus on performance improvements have not shown an overwhelming performance compared with existing models. In this study, we propose MonaCoBERT, which achieves the best performance on most benchmark datasets and has significant interpretability. MonaCoBERT uses a BERT-based architecture with monotonic convolutional multihead attention, which reflects forgetting behavior of the students and increases the representation power of the model. We can also increase the performance and interpretability using a classical test-theory-based (CTT-based) embedding strategy that considers the difficulty of the question. To determine why MonaCoBERT achieved the best performance and interpret the results quantitatively, we conducted ablation studies and additional analyses using Grad-CAM, UMAP, and various visualization techniques. The analysis results demonstrate that both attention components complement one another and that CTT-based embedding represents information on both global and local difficulties. We also demonstrate that our model represents the relationship between concepts.

cs.CY↗

arXiv2022

How to Fine-tune Models with Few Samples: Update, Data Augmentation, and Test-time Augmentation

Most of the recent few-shot learning (FSL) algorithms are based on transfer learning, where a model is pre-trained using a large amount of source data, and the pre-trained model is fine-tuned using a small amount of target data. In transfer learning-based FSL, sophisticated pre-training methods have been widely studied for universal representation. Therefore, it has become more important to utilize the universal representation for downstream tasks, but there are few studies on fine-tuning in FSL. In this paper, we focus on how to transfer pre-trained models to few-shot downstream tasks from the three perspectives: update, data augmentation, and test-time augmentation. First, we compare the two popular update methods, full fine-tuning (i.e., updating the entire network, FT) and linear probing (i.e., updating only a linear classifier, LP). We find that LP is better than FT with extremely few samples, whereas FT outperforms LP as training samples increase. Next, we show that data augmentation cannot guarantee few-shot performance improvement and investigate the effectiveness of data augmentation based on the intensity of augmentation. Finally, we adopt augmentation to both a support set for update (i.e., data augmentation) as well as a query set for prediction (i.e., test-time augmentation), considering support-query distribution shifts, and improve few-shot performance. The code is available at https://github.com/kimyuji/updating_FSL.

cs.LG↗

arXiv2022

There is no rose without a thorn: Finding weaknesses on BlenderBot 2.0 in terms of Model, Data and User-Centric Approach

BlenderBot 2.0 is a dialogue model that represents open-domain chatbots by reflecting real-time information and remembering user information for an extended period using an internet search module and multi-session. Nonetheless, the model still has room for improvement. To this end, we examine BlenderBot 2.0 limitations and errors from three perspectives: model, data, and user. From the data point of view, we highlight the unclear guidelines provided to workers during the crowdsourcing process, as well as a lack of a process for refining hate speech in the collected data and verifying the accuracy of internet-based information. From a user perspective, we identify nine types of limitations of BlenderBot 2.0, and their causes are thoroughly investigated. Furthermore, for each point of view, we propose practical improvement methods and discuss several potential future research directions.

cs.CL↗