arXiv subjects

Chako Takahashi

Publications and source records attributed to Chako Takahashi.

7 recordsLinked to original sources

arXiv2026

Temporal Portability of Numeric User Metadata on Twitter

Numeric user metadata in social media are often reused over time. However, their reusability may depend on what an analysis needs to preserve. We introduce temporal portability as an analytical perspective for assessing the cross-time reuse of user features and feature-based rules. Specifically, we ask how well relevant properties are preserved when features and rules defined at a source time point are reused at a target time point. We used quarterly data on user features obtained directly from or derived from Japanese-language tweets in Twitter's 1% sample stream from 2020-Q1 to 2022-Q3. Each quarter included approximately 10.1--11.0 million unique users. We evaluated 13 numeric user features in terms of feature distributions, same-user relative ranks, selection rates, and selected-user membership. Across quarters, feature distributions changed and, for many features, same-user relative ranks were less well preserved at longer quarter lags. Reusing source-quarter thresholds also produced selection-rate drift. Target-quarter recalibration nearly matched source-quarter selection rates. However, membership turnover persisted and increased at longer quarter lags. Our results show that temporal portability should be assessed in terms of the property that an analysis needs to preserve.

cs.SI↗

arXiv2025

Dataset-Free Weight-Initialization on Restricted Boltzmann Machine

In feed-forward neural networks, dataset-free weight-initialization methods such as LeCun, Xavier (or Glorot), and He initializations have been developed. These methods randomly determine the initial values of weight parameters based on specific distributions (e.g., Gaussian or uniform distributions) without using training datasets. To the best of the authors' knowledge, such a dataset-free weight-initialization method is yet to be developed for restricted Boltzmann machines (RBMs), which are probabilistic neural networks consisting of two layers. In this study, we derive a dataset-free weight-initialization method for Bernoulli--Bernoulli RBMs based on statistical mechanical analysis. In the proposed weight-initialization method, the weight parameters are drawn from a Gaussian distribution with zero mean. The standard deviation of the Gaussian distribution is optimized based on our hypothesis that a standard deviation providing a larger layer correlation (LC) between the two layers improves the learning efficiency. The expression of the LC is derived based on a statistical mechanical analysis. The optimal value of the standard deviation corresponds to the maximum point of the LC. The proposed weight-initialization method is identical to Xavier initialization in a specific case (i.e., when the sizes of the two layers are the same, the random variables of the layers are $\{-1,1\}$-binary, and all bias parameters are zero). The validity of the proposed weight-initialization method is demonstrated in numerical experiments using a toy and real-world datasets.

stat.ML↗

arXiv2022

Free Energy Evaluation Using Marginalized Annealed Importance Sampling

The evaluation of the free energy of a stochastic model is considered a significant issue in various fields of physics and machine learning. However, the exact free energy evaluation is computationally infeasible because the free energy expression includes an intractable partition function. Annealed importance sampling (AIS) is a type of importance sampling based on the Markov chain Monte Carlo method that is similar to a simulated annealing and can effectively approximate the free energy. This study proposes an AIS-based approach, which is referred to as marginalized AIS (mAIS). The statistical efficiency of mAIS is investigated in detail based on theoretical and numerical perspectives. Based on the investigation, it is proved that mAIS is more effective than AIS under a certain condition.

stat.ML↗

arXiv2020

Adaptive Thouless--Anderson--Palmer equation for higher-order Markov random fields

The adaptive Thouless--Anderson--Palmer (TAP) mean-field approximation is one of the advanced mean-field approaches, and it is known as a powerful accurate method for Markov random fields (MRFs) with quadratic interactions (pairwise MRFs). In this study, an extension of the adaptive TAP approximation for MRFs with many-body interactions (higher-order MRFs) is developed. We show that the adaptive TAP equation for pairwise MRFs is derived by naive mean-field approximation with diagonal consistency. Based on the equivalence of the approximate equation obtained from the naive mean-field approximation with diagonal consistency and the adaptive TAP equation in pairwise MRFs, we formulate approximate equations for higher-order Boltzmann machines, which is one of simplest higher-order MRFs, via the naive mean-field approximation with diagonal consistency.

cond-mat.dis-nn↗

arXiv2018

Perturbative Interpretation of Adaptive Thouless-Anderson-Palmer Free Energy

In conventional well-known derivation methods for the adaptive Thouless-Anderson-Palmer (TAP) free energy, special assumptions that are difficult to mathematically justify except in some mean-field models, must be made. Here, we present a new adaptive TAP free energy derivation method. Using this derivation technique, without any special assumptions, the adaptive TAP free energy can be simply obtained as a high-temperature expansion of the Gibbs free energy.

cond-mat.dis-nn↗

arXiv2018

Statistical-mechanical analysis of compressed sensing for Hamiltonian estimation of Ising spin glass

Several powerful machines, such as the D-Wave 2000Q, dedicated to solving combinatorial optimization problems through the Ising-model formulation have been developed. To input problems into the machines, the unknown parameters on the Ising model must be determined, and this is necessarily a nontrivial task. It could be beneficial to construct a method to estimate the parameters of the Ising model from several pairs of values of the energy and spin configurations. In the present paper, we propose a simple method employing the $L_1$-norm minimization, which is based on the concept of the compressed sensing. Moreover, we analyze the typical performance of our proposed method of the Hamiltonian estimation by using the replica method. We also compare our analytical results through several numerical experiments using the alternating direction method of multipliers.

cond-mat.stat-mech↗

arXiv2016

Mean-Field Inference in Gaussian Restricted Boltzmann Machine

A Gaussian restricted Boltzmann machine (GRBM) is a Boltzmann machine defined on a bipartite graph and is an extension of usual restricted Boltzmann machines. A GRBM consists of two different layers: a visible layer composed of continuous visible variables and a hidden layer composed of discrete hidden variables. In this paper, we derive two different inference algorithms for GRBMs based on the naive mean-field approximation (NMFA). One is an inference algorithm for whole variables in a GRBM, and the other is an inference algorithm for partial variables in a GBRBM. We compare the two methods analytically and numerically and show that the latter method is better.

stat.ML↗