Picture for Beomsu Kim

Beomsu Kim

Bridging Active Exploration and Uncertainty-Aware Deployment Using Probabilistic Ensemble Neural Network Dynamics

May 20, 2023
Taekyung Kim, Jungwi Mun, Junwon Seo, Beomsu Kim, Seongil Hong

* 2023 Robotics: Science and Systems (RSS). Project page: https://taekyung.me/rss2023-bridging

Access Paper or Ask Questions

Measuring and Improving Semantic Diversity of Dialogue Generation

Oct 11, 2022
Seungju Han, Beomsu Kim, Buru Chang

Figure 1 for Measuring and Improving Semantic Diversity of Dialogue Generation

Figure 2 for Measuring and Improving Semantic Diversity of Dialogue Generation

Figure 3 for Measuring and Improving Semantic Diversity of Dialogue Generation

Figure 4 for Measuring and Improving Semantic Diversity of Dialogue Generation

* EMNLP22 Findings

Access Paper or Ask Questions

Denoising MCMC for Accelerating Diffusion-Based Generative Models

Sep 29, 2022
Beomsu Kim, Jong Chul Ye

Figure 1 for Denoising MCMC for Accelerating Diffusion-Based Generative Models

Figure 2 for Denoising MCMC for Accelerating Diffusion-Based Generative Models

Figure 3 for Denoising MCMC for Accelerating Diffusion-Based Generative Models

Figure 4 for Denoising MCMC for Accelerating Diffusion-Based Generative Models

Access Paper or Ask Questions

Mitigating Out-of-Distribution Data Density Overestimation in Energy-Based Models

May 30, 2022
Beomsu Kim, Jong Chul Ye

Figure 1 for Mitigating Out-of-Distribution Data Density Overestimation in Energy-Based Models

Figure 2 for Mitigating Out-of-Distribution Data Density Overestimation in Energy-Based Models

Figure 3 for Mitigating Out-of-Distribution Data Density Overestimation in Energy-Based Models

Figure 4 for Mitigating Out-of-Distribution Data Density Overestimation in Energy-Based Models

Access Paper or Ask Questions

Semi-Implicit Hybrid Gradient Methods with Application to Adversarial Robustness

Feb 21, 2022
Beomsu Kim, Junghoon Seo

Figure 1 for Semi-Implicit Hybrid Gradient Methods with Application to Adversarial Robustness

Figure 2 for Semi-Implicit Hybrid Gradient Methods with Application to Adversarial Robustness

Figure 3 for Semi-Implicit Hybrid Gradient Methods with Application to Adversarial Robustness

Figure 4 for Semi-Implicit Hybrid Gradient Methods with Application to Adversarial Robustness

* International Conference on Artificial Intelligence and Statistics (AISTATS) 2022

Access Paper or Ask Questions

Energy-Based Contrastive Learning of Visual Representations

Feb 10, 2022
Beomsu Kim, Jong Chul Ye

Figure 1 for Energy-Based Contrastive Learning of Visual Representations

Figure 2 for Energy-Based Contrastive Learning of Visual Representations

Figure 3 for Energy-Based Contrastive Learning of Visual Representations

Figure 4 for Energy-Based Contrastive Learning of Visual Representations

Access Paper or Ask Questions