Search

The formation mechanics of 1.7 turn of superhelix of DNA strands from sequence dependent nonlinear energetics including bend twist coupling is the quintessential step of DNA packaging that is remained elusive. In this paper, the geometrical modeling of each base pair in the curved DNA strand addresses the sequence dependent deformation energy during a superhelix formation around a simplified core structure. The quantization of the energy involved in the bend-twist coupling deformation in base pair wise unit characterizes the accessible path for a superhelix deformation. Coarse-grained molecular dynamic simulation validates the description of the curvature formation process which overcomes its persistence length and the mechanics involved in 1.7 turn of superhelix curvature.

The Scalar Field Dark Matter model has been known in various ways throughout its history; Fuzzy, BEC, Wave, Ultralight, Axion-like Dark Matter, etc. All of them consist in proposing that the dark matter of the universe is a spinless field $\Phi$ that follows the Klein-Gordon (KG) equation of motion $\Box\Phi-dV/d\Phi=0$, for a given scalar field potential $V$. The difference between different models is sometimes the choice of the scalar field potential $V$. In the literature we find that people usually work in the nonrelativistic, weak-field limit of the KG equation where it transforms into the Schr\"odinger equation and the Einstein equations into the Poisson equation, reducing the KG-Einstein system, to the Schr\"odinger-Poisson system. In this paper, we review some of the most interesting achievements of this model from the historical point of view and its comparison with observations, showing that this model could be the last answer to the question about the nature of dark matter in the universe., Comment: 8 pages, no figures

Previous logits-based Knowledge Distillation (KD) have utilized predictions about multiple categories within each sample (i.e., class predictions) and have employed Kullback-Leibler (KL) divergence to reduce the discrepancy between the student and teacher predictions. Despite the proliferation of KD techniques, the student model continues to fall short of achieving a similar level as teachers. In response, we introduce a novel and effective KD method capable of achieving results on par with or superior to the teacher models performance. We utilize teacher and student predictions about multiple samples for each category (i.e., batch predictions) and apply cosine similarity, a commonly used technique in Natural Language Processing (NLP) for measuring the resemblance between text embeddings. This metric's inherent scale-invariance property, which relies solely on vector direction and not magnitude, allows the student to dynamically learn from the teacher's knowledge, rather than being bound by a fixed distribution of the teacher's knowledge. Furthermore, we propose a method called cosine similarity weighted temperature (CSWT) to improve the performance. CSWT reduces the temperature scaling in KD when the cosine similarity between the student and teacher models is high, and conversely, it increases the temperature scaling when the cosine similarity is low. This adjustment optimizes the transfer of information from the teacher to the student model. Extensive experimental results show that our proposed method serves as a viable alternative to existing methods. We anticipate that this approach will offer valuable insights for future research on model compression., Comment: 13 pages, 5 figures

Pre-trained Vision Language Models (VLMs) have demonstrated notable progress in various zero-shot tasks, such as classification and retrieval. Despite their performance, because improving performance on new tasks requires task-specific knowledge, their adaptation is essential. While labels are needed for the adaptation, acquiring them is typically expensive. To overcome this challenge, active learning, a method of achieving a high performance by obtaining labels for a small number of samples from experts, has been studied. Active learning primarily focuses on selecting unlabeled samples for labeling and leveraging them to train models. In this study, we pose the question, "how can the pre-trained VLMs be adapted under the active learning framework?" In response to this inquiry, we observe that (1) simply applying a conventional active learning framework to pre-trained VLMs even may degrade performance compared to random selection because of the class imbalance in labeling candidates, and (2) the knowledge of VLMs can provide hints for achieving the balance before labeling. Based on these observations, we devise a novel active learning framework for VLMs, denoted as PCB. To assess the effectiveness of our approach, we conduct experiments on seven different real-world datasets, and the results demonstrate that PCB surpasses conventional active learning and random sampling methods. Code will be available in https://github.com/kaist-dmlab/pcb ., Comment: accepted at CVPR 2024

In real-world continual learning (CL) scenarios, tasks often exhibit intricate and unpredictable semantic shifts, posing challenges for fixed prompt management strategies which are tailored to only handle semantic shifts of uniform degree (i.e., uniformly mild or uniformly abrupt). To address this limitation, we propose an adaptive prompting approach that effectively accommodates semantic shifts of varying degree where mild and abrupt shifts are mixed. AdaPromptCL employs the assign-and-refine semantic grouping mechanism that dynamically manages prompt groups in accordance with the semantic similarity between tasks, enhancing the quality of grouping through continuous refinement. Our experiment results demonstrate that AdaPromptCL outperforms existing prompting methods by up to 21.3%, especially in the benchmark datasets with diverse semantic shifts between tasks., Comment: ICML 2024

When two galaxies merge, they often produce a supermassive black hole binary (SMBHB) at their center. Numerical simulations with stars and cold dark matter show that SMBHBs typically stall out at a distance of a few parsecs apart and take billions of years to coalesce. This is known as the final parsec problem. We suggest that ultralight dark matter (ULDM) halos around SMBHBs can generate dark matter waves due to dynamical friction. These waves can effectively carry away orbital energy from the black holes, rapidly driving them together. To test this hypothesis, we performed numerical simulations of black hole binaries inside ULDM halos. Due to gravitational cooling and quasi-normal modes, the loss-cone problem can be avoided. The decay time scale gives lower bounds on masses of the ULDM particles and SMBHBs comparable to observational data. Our results imply that ULDM waves can lead to the rapid orbital decay of black hole binaries., Comment: Accepted for publication in Physics Letters B

Recent years have witnessed a surge of discoveries in the studies of thermodynamic inequalities: the thermodynamic uncertainty relation (TUR) and the entropic bound (EB) provide a lower bound on the entropy production (EP) in terms of nonequilibrium currents; the classical speed limit (CSL) expresses the lower bound on the EP using the geometry of probability distributions; the power-efficiency (PE) tradeoff dictates the maximum power achievable for a heat engine given the level of its thermal efficiency. In this study, we show that there exists a unified hierarchical structure encompassing all of these bounds, with the fundamental inequality given by a novel extension of the TUR (XTUR) that incorporates the most general range of current-like and state-dependent observables. By selecting more specific observables, the TUR and the EB follow from the XTUR, and the CSL and the PE tradeoff follow from the EB. Our derivations cover both Langevin and Markov jump systems, with the first proof of the EB for the Markov jump systems and a more generalized form of the CSL. We also present concrete examples of the EB for the Markov jump systems and the generalized CSL., Comment: 23 pages, 5 figures