Searcharxiv⌕ Search

arXiv subjects

Takashi Sano

Publications and source records attributed to Takashi Sano.

10 recordsLinked to original sources

Physics2Weight: Direct Neural Network Architecture and Weight Generation from Physical Laws Using Large Language Models

Recent advances in large language models (LLMs) have enabled weight-space programming (WSP), which directly generates neural network weights without gradient-based optimization. However, extending WSP to physical problems remains challenging, particularly in achieving both solution accuracy and physical consistency. In this paper, we propose Physics2Weight (PtW), a framework that extends WSP to solving partial differential equations by directly generating neural network weights and editing architectures based on governing equations and initial and boundary conditions. PtW introduces physics-constrained prompting and physics-guided architecture editing, combined with iterative weight refinement from WSP, to improve solution accuracy and physical consistency without gradient-based optimization. Experiments on four PDEs, namely the heat, advection, Burgers', and Allen--Cahn equations, demonstrate lower solution errors than conventional physics-informed neural networks, while revealing limitations in satisfying the governing equations for nonlinear PDEs.

cs.AI↗

Calibrating Track-Query Updates at Test Time for Multi-Object Tracking

Multi-Object Tracking (MOT) requires localizing multiple objects and maintaining their identities over time. Recent query-propagation trackers carry target-specific representations across frames and update them with current-frame image features, yet identity association remains difficult when visually similar targets overlap or cross. During such interactions, current-frame query updates may contain components directed toward nearby active tracks, which may affect identity association in later frames. To investigate this issue, we analyze track-query updates preceding identity switches (IDSWs), where the same target is assigned a different track ID. Our analysis shows that (i) propagated track-query representations change more abruptly before IDSWs, (ii) current-frame updates have larger positive projections along directions toward other active tracks, and (iii) bounding-box intersection over union (IoU) retrieves the counterpart track involved in a subsequent IDSW with high recall when that track remains active before the switch. Motivated by these findings, we propose Test-time Track-Query Calibration (TQC), a training-free and gradient-free method for query-propagation trackers. TQC calibrates the tracker output before propagation using a bounded correction estimated from update components directed toward nearby active tracks. It requires only track-query representations and bounding boxes available during inference, without retraining or architectural modification. Across nine DanceTrack and SportsMOT test settings, TQC improves HOTA and AssA in all settings and IDF1 in eight, with an average HOTA gain of +0.38 and limited inference overhead. Controlled variants on six validation settings further show that TQC's correction direction yields more consistent HOTA improvements than alternative directions with matched correction norms.

cs.CV↗

Investigation of Frame Differences as Motion Cues for Video Object Segmentation

Automatic Video Object Segmentation (AVOS) refers to the task of autonomously segmenting target objects in video sequences without relying on human-provided annotations in the first frames. In AVOS, the use of motion information is crucial, with optical flow being a commonly employed method for capturing motion cues. However, the computation of optical flow is resource-intensive, making it unsuitable for real-time applications, especially on edge devices with limited computational resources. In this study, we propose using frame differences as an alternative to optical flow for motion cue extraction. We developed an extended U-Net-like AVOS model that takes a frame on which segmentation is performed and a frame difference as inputs, and outputs an estimated segmentation map. Our experimental results demonstrate that the proposed model achieves performance comparable to the model with optical flow as an input, particularly when applied to videos captured by stationary cameras. Our results suggest the usefulness of employing frame differences as motion cues in cases with limited computational resources.

cs.CV↗

From Coupled Oscillators to Graph Neural Networks: Reducing Over-smoothing via a Kuramoto Model-based Approach

We propose the Kuramoto Graph Neural Network (KuramotoGNN), a novel class of continuous-depth graph neural networks (GNNs) that employs the Kuramoto model to mitigate the over-smoothing phenomenon, in which node features in GNNs become indistinguishable as the number of layers increases. The Kuramoto model captures the synchronization behavior of non-linear coupled oscillators. Under the view of coupled oscillators, we first show the connection between Kuramoto model and basic GNN and then over-smoothing phenomenon in GNNs can be interpreted as phase synchronization in Kuramoto model. The KuramotoGNN replaces this phase synchronization with frequency synchronization to prevent the node features from converging into each other while allowing the system to reach a stable synchronized state. We experimentally verify the advantages of the KuramotoGNN over the baseline GNNs and existing methods in reducing over-smoothing on various graph deep learning benchmark tasks.

cs.LG↗

An Analytic Solution to the Inverse Ising Problem in the Tree-reweighted Approximation

Many iterative and non-iterative methods have been developed for inverse problems associated with Ising models. Aiming to derive an accurate non-iterative method for the inverse problems, we employ the tree-reweighted approximation. Using the tree-reweighted approximation, we can optimize the rigorous lower bound of the objective function. By solving the moment-matching and self-consistency conditions analytically, we can derive the interaction matrix as a function of the given data statistics. With this solution, we can obtain the optimal interaction matrix without iterative computation. To evaluate the accuracy of the proposed inverse formula, we compared our results to those obtained by existing inverse formulae derived with other approximations. In an experiment to reconstruct the interaction matrix, we found that the proposed formula returns the best estimates in strongly-attractive regions for various graph structures. We also performed an experiment using real-world biological data. When applied to finding the connectivity of neurons from spike train data, the proposed formula gave the closest result to that obtained by a gradient ascent algorithm, which typically requires thousands of iterations.

stat.ML↗

Random matrix model for chiral and color-flavor locking condensates

We study the phase diagram of a chiral random matrix model with three quark flavors at finite temperature and chemical potential, taking the chiral and diquark condensates as independent order parameters. Fixing the ratio of the coupling strengths in the quark-antiquark and quark-quark channels applying the Fierz transformation, we find that the color-flavor locked (CFL) phase is realized at large chemical potential, while the ordinary chirally-broken phase appears in the region with small chemical potential. We investigate responses of the phases by changing small quark masses in the cases with three equal-mass flavors and with 2+1 flavors. In the case with three equal-mass flavors, we find that the finite masses make the CFL phase transition line move to the higher density region. In the case with 2+1 flavors, we find the two-flavor color superconducting phase at the medium density region as a result of the finite asymmetry between the flavors, as well as the CFL phase at higher density region.

hep-ph↗

Loewner matrices of matrix convex and monotone functions

The matrix convexity and the matrix monotony of a real $C^1$ function $f$ on $(0,\infty)$ are characterized in terms of the conditional negative or positive definiteness of the Loewner matrices associated with $f$, $tf(t)$, and $t^2f(t)$. Similar characterizations are also obtained for matrix monotone functions on a finite interval $(a,b)$.

math.FA↗

Chiral phase transition in a random matrix model with three flavors

The chiral phase transition in the conventional random matrix model is the second order in the chiral limit, irrespective of the number of flavors N_f, because it lacks the U_A(1)-breaking determinant interaction term. Furthermore, it predicts an unphysical value of zero for the topological susceptibility at finite temperatures. We propose a new chiral random matrix model which resolves these difficulties by incorporating the determinant interaction term within the instanton gas picture. The model produces a second-order transition for N_f=2 and a first-order transition for N_f=3, and recovers a physical temperature dependence of the topological susceptibility.

hep-lat↗

The Ashtekar Formalism and WKB Wave Functions of N=1,2 Supergravities

The N=1,2 supergravities with non-zero cosmological constants are investigated in the Ashtekar formalism. We solve the constraints of the N=1,2 supergravities semi-classically. The resulting WKB wave functions are expressed by exponentials of supersymmetric-extended SL(2,C) Chern-Simons functional.

hep-th↗

The Non-perturbative Canonical Quantization of the N=1 Supergravity

The non-perturbative canonical quantization of the N=1 supergravity with the non-zero cosmological constant is studied using the Ashtekar formalism. A semi-classical wave function is obtained and it has the form of the exponential of the N=1 supersymmetric extension of the Chern-Simons functional. The N=1 supergravity in the Robertson-Walker universe is also examined and some analytic solutions are obtained. (We omit all graphs in the following LaTex file.)

hep-th↗