Jihoon Tack
YOU?
Author Swipe
View article: Think Clearly: Improving Reasoning via Redundant Token Pruning
Think Clearly: Improving Reasoning via Redundant Token Pruning Open
Recent large language models have shown promising capabilities in long-form reasoning, following structured chains of thought before arriving at a final answer. However, we observe that these reasoning paths tend to include substantial red…
View article: Mamba Drafters for Speculative Decoding
Mamba Drafters for Speculative Decoding Open
Speculative decoding has emerged as a promising approach to accelerating large language model (LLM) generation using a fast drafter while maintaining alignment with the target model's distribution. However, existing approaches face a trade…
View article: ReGUIDE: Data Efficient GUI Grounding via Spatial Reasoning and Search
ReGUIDE: Data Efficient GUI Grounding via Spatial Reasoning and Search Open
Recent advances in Multimodal Large Language Models (MLLMs) have enabled autonomous agents to interact with computers via Graphical User Interfaces (GUIs), where accurately localizing the coordinates of interface elements (e.g., buttons) i…
View article: Adversarial Self-Supervised Contrastive Learning
Adversarial Self-Supervised Contrastive Learning Open
Existing adversarial learning approaches mostly use class labels to generate adversarial samples that lead to incorrect predictions, which are then used to augment the training of the model for improved robustness. While some recent works …
View article: Tabular Transfer Learning via Prompting LLMs
Tabular Transfer Learning via Prompting LLMs Open
Learning with a limited number of labeled data is a central problem in real-world applications of machine learning, as it is often expensive to obtain annotations. To deal with the scarcity of labeled data, transfer learning is a conventio…
View article: Optimized Feature Generation for Tabular Data via LLMs with Decision Tree Reasoning
Optimized Feature Generation for Tabular Data via LLMs with Decision Tree Reasoning Open
In tabular prediction tasks, tree-based models combined with automated feature engineering methods often outperform deep learning approaches that rely on learned representations. While these feature engineering techniques are effective, th…
View article: Unleashing the Power of Meta-tuning for Few-shot Generalization Through Sparse Interpolated Experts
Unleashing the Power of Meta-tuning for Few-shot Generalization Through Sparse Interpolated Experts Open
Recent successes suggest that parameter-efficient fine-tuning of foundation models as the state-of-the-art method for transfer learning in vision, replacing the rich literature of alternatives such as meta-learning. In trying to harness th…
View article: Online Adaptation of Language Models with a Memory of Amortized Contexts
Online Adaptation of Language Models with a Memory of Amortized Contexts Open
Due to the rapid generation and dissemination of information, large language models (LLMs) quickly run out of date despite enormous development costs. To address the crucial need to keep models updated, online learning has emerged as a cri…
View article: Modality-Agnostic Self-Supervised Learning with Meta-Learned Masked Auto-Encoder
Modality-Agnostic Self-Supervised Learning with Meta-Learned Masked Auto-Encoder Open
Despite its practical importance across a wide range of modalities, recent advances in self-supervised learning (SSL) have been primarily focused on a few well-curated domains, e.g., vision and language, often relying on their domain-speci…
View article: STUNT: Few-shot Tabular Learning with Self-generated Tasks from Unlabeled Tables
STUNT: Few-shot Tabular Learning with Self-generated Tasks from Unlabeled Tables Open
Learning with few labeled tabular samples is often an essential requirement for industrial machine learning applications as varieties of tabular data suffer from high annotation costs or have difficulties in collecting new samples for nove…
View article: Learning Large-scale Neural Fields via Context Pruned Meta-Learning
Learning Large-scale Neural Fields via Context Pruned Meta-Learning Open
We introduce an efficient optimization-based meta-learning technique for large-scale neural field training by realizing significant memory savings through automated online context point selection. This is achieved by focusing each learning…
View article: Modality-Agnostic Variational Compression of Implicit Neural Representations
Modality-Agnostic Variational Compression of Implicit Neural Representations Open
We introduce a modality-agnostic neural compression algorithm based on a functional view of data and parameterised as an Implicit Neural Representation (INR). Bridging the gap between latent coding and sparsity, we obtain compact latent re…
View article: Meta-Learning with Self-Improving Momentum Target
Meta-Learning with Self-Improving Momentum Target Open
The idea of using a separately trained target model (or teacher) to improve the performance of the student model has been increasingly popular in various machine learning domains, and meta-learning is no exception; a recent discovery shows…
View article: Consistency Regularization for Adversarial Robustness
Consistency Regularization for Adversarial Robustness Open
Adversarial training (AT) is currently one of the most successful methods to obtain the adversarial robustness of deep neural networks. However, the phenomenon of robust overfitting, i.e., the robustness starts to decrease significantly du…
View article: Generating Videos with Dynamics-aware Implicit Generative Adversarial Networks
Generating Videos with Dynamics-aware Implicit Generative Adversarial Networks Open
In the deep learning era, long video generation of high-quality still remains challenging due to the spatio-temporal complexity and continuity of videos. Existing prior works have attempted to model video distribution by representing video…
View article: Meta-Learning Sparse Implicit Neural Representations
Meta-Learning Sparse Implicit Neural Representations Open
Implicit neural representations are a promising new avenue of representing general signals by learning a continuous function that, parameterized as a neural network, maps the domain of a signal to its codomain; the mapping from spatial coo…
View article: Consistency Regularization for Adversarial Robustness
Consistency Regularization for Adversarial Robustness Open
Adversarial training (AT) is currently one of the most successful methods to obtain the adversarial robustness of deep neural networks. However, the phenomenon of robust overfitting, i.e., the robustness starts to decrease significantly du…
View article: CSI: Novelty Detection via Contrastive Learning on Distributionally Shifted Instances
CSI: Novelty Detection via Contrastive Learning on Distributionally Shifted Instances Open
Novelty detection, i.e., identifying whether a given sample is drawn from outside the training distribution, is essential for reliable machine learning. To this end, there have been many attempts at learning a representation well-suited fo…
View article: CSI: Novelty Detection via Contrastive Learning on Distributionally Shifted Instances
CSI: Novelty Detection via Contrastive Learning on Distributionally Shifted Instances Open
Novelty detection, i.e., identifying whether a given sample is drawn from outside the training distribution, is essential for reliable machine learning. To this end, there have been many attempts at learning a representation well-suited fo…
View article: Adversarial Self-Supervised Contrastive Learning
Adversarial Self-Supervised Contrastive Learning Open
Existing adversarial learning approaches mostly use class labels to generate adversarial samples that lead to incorrect predictions, which are then used to augment the training of the model for improved robustness. While some recent works …