arXiv subjects

Makoto Takamoto

Publications and source records attributed to Makoto Takamoto.

At least 19 recordsLinked to original sources

arXiv2026

On the limits and opportunities of AI reviewers: Reviewing the reviews of Nature-family papers with 45 expert scientists

With the advancement of AI capabilities, AI reviewers are beginning to be deployed in scientific peer review, yet their capability and credibility remain in question: many scientists simply view them as probabilistic systems without the expertise to evaluate research, while other researchers are more optimistic about their readiness without concrete evidence. Understanding what AI reviewers do well, where they fall short, and what challenges remain is essential. However, existing evaluations of AI reviewers have focused on whether their verdicts match human verdicts (e.g., score alignment, acceptance prediction), which is insufficient to characterize their capabilities and limits. In this paper, we close this gap through a large-scale expert annotation study, in which 45 domain scientists in Physical, Biological, and Health Sciences spent 469 hours rating 2,960 individual criticisms (each targeting one specific aspect of a paper) from human-written and AI-generated reviews of 82 Nature-family papers on correctness, significance, and sufficiency of evidence. On a composite of all three dimensions, a reviewing agent powered by GPT-5.2 scores above each paper's top-rated human reviewer (60.0% vs. 48.2%, p = 0.009), while all three AI reviewers (including Gemini 3.0 Pro and Claude Opus 4.5) exceed the lowest-rated human across every dimension. AI reviewers' accurate criticisms are also more often rated significant and well-evidenced, and surface a distinct 26% of issues no human raises. However, AI reviewers overlap far more than humans do (21% vs. 3% for cross-reviewer pairs), and exhibit 16 recurring weaknesses humans do not share, such as limited subfield knowledge, lack of long context management over multiple files, and overly critical stance on minor issues. Overall, our results position current AI reviewers as complements to, not substitutes for, human reviewers.

cs.CL↗

arXiv2025

Fast, Modular, and Differentiable Framework for Machine Learning-Enhanced Molecular Simulations

We present an end-to-end differentiable molecular simulation framework (DIMOS) for molecular dynamics and Monte Carlo simulations. DIMOS easily integrates machine-learning-based interatomic potentials and implements classical force fields including an efficient implementation of particle-mesh Ewald. Thanks to its modularity, both classical and machine-learning-based approaches can be easily combined into a hybrid description of the system (ML/MM). By supporting key molecular dynamics features such as efficient neighborlists and constraint algorithms for larger time steps, the framework makes steps in bridging the gap between hand-optimized simulation engines and the flexibility of a \verb|PyTorch| implementation. We show that due to improved linear instead of quadratic scaling as function of system size DIMOS is able to obtain speed-up factors of up to $170\times$ for classical force field simulations against another fully differentiable simulation framework. The advantage of differentiability is demonstrated by an end-to-end optimization of the proposal distribution in a Markov Chain Monte Carlo simulation based on Hamiltonian Monte Carlo (HMC). Using these optimized simulation parameters a $3\times$ acceleration is observed in comparison to ad-hoc chosen simulation parameters. The code is available at https://github.com/nec-research/DIMOS.

physics.comp-ph↗

arXiv2025

Physics-Informed Weakly Supervised Learning for Interatomic Potentials

Machine learning plays an increasingly important role in computational chemistry and materials science, complementing computationally intensive ab initio and first-principles methods. Despite their utility, machine-learning models often lack generalization capability and robustness during atomistic simulations, yielding unphysical energy and force predictions that hinder their real-world applications. We address this challenge by introducing a physics-informed, weakly supervised approach for training machine-learned interatomic potentials (MLIPs). We introduce two novel loss functions, extrapolating the potential energy via a Taylor expansion and using the concept of conservative forces. Our approach improves the accuracy of MLIPs applied to training tasks with sparse training data sets and reduces the need for pre-training computationally demanding models with large data sets. Particularly, we perform extensive experiments demonstrating reduced energy and force errors -- often lower by a factor of two -- for various baseline models and benchmark data sets. Moreover, we demonstrate improved robustness during MD simulations of the MLIP models trained with the proposed weakly supervised loss. Finally, our approach improves the fine-tuning of foundation models on sparse, highly accurate ab initio data. An implementation of our method and scripts for executing experiments are available at https://github.com/nec-research/PICPS-ML4Sci.

physics.chem-ph↗

arXiv2025

Optimal Embedding Guided Negative Sample Generation for Knowledge Graph Link Prediction

Knowledge graph embedding (KGE) models encode the structural information of knowledge graphs to predicting new links. Effective training of these models requires distinguishing between positive and negative samples with high precision. Although prior research has shown that improving the quality of negative samples can significantly enhance model accuracy, identifying high-quality negative samples remains a challenging problem. This paper theoretically investigates the condition under which negative samples lead to optimal KG embedding and identifies a sufficient condition for an effective negative sample distribution. Based on this theoretical foundation, we propose \textbf{E}mbedding \textbf{MU}tation (\textsc{EMU}), a novel framework that \emph{generates} negative samples satisfying this condition, in contrast to conventional methods that focus on \emph{identifying} challenging negative samples within the training data. Importantly, the simplicity of \textsc{EMU} ensures seamless integration with existing KGE models and negative sampling methods. To evaluate its efficacy, we conducted comprehensive experiments across multiple datasets. The results consistently demonstrate significant improvements in link prediction performance across various KGE models and negative sampling methods. Notably, \textsc{EMU} enables performance improvements comparable to those achieved by models with embedding dimension five times larger. An implementation of the method and experiments are available at https://github.com/nec-research/EMU-KG.

cs.LG↗

arXiv2025

Active Learning for Neural PDE Solvers

Solving partial differential equations (PDEs) is a fundamental problem in science and engineering. While neural PDE solvers can be more efficient than established numerical solvers, they often require large amounts of training data that is costly to obtain. Active learning (AL) could help surrogate models reach the same accuracy with smaller training sets by querying classical solvers with more informative initial conditions and PDE parameters. While AL is more common in other domains, it has yet to be studied extensively for neural PDE solvers. To bridge this gap, we introduce AL4PDE, a modular and extensible active learning benchmark. It provides multiple parametric PDEs and state-of-the-art surrogate models for the solver-in-the-loop setting, enabling the evaluation of existing and the development of new AL methods for neural PDE solving. We use the benchmark to evaluate batch active learning algorithms such as uncertainty- and feature-based methods. We show that AL reduces the average error by up to 71% compared to random sampling and significantly reduces worst-case errors. Moreover, AL generates similar datasets across repeated runs, with consistent distributions over the PDE parameters and initial conditions. The acquired datasets are reusable, providing benefits for surrogate models not involved in the data generation.

cs.LG↗

arXiv2024

Higher-Rank Irreducible Cartesian Tensors for Equivariant Message Passing

The ability to perform fast and accurate atomistic simulations is crucial for advancing the chemical sciences. By learning from high-quality data, machine-learned interatomic potentials achieve accuracy on par with ab initio and first-principles methods at a fraction of their computational cost. The success of machine-learned interatomic potentials arises from integrating inductive biases such as equivariance to group actions on an atomic system, e.g., equivariance to rotations and reflections. In particular, the field has notably advanced with the emergence of equivariant message passing. Most of these models represent an atomic system using spherical tensors, tensor products of which require complicated numerical coefficients and can be computationally demanding. Cartesian tensors offer a promising alternative, though state-of-the-art methods lack flexibility in message-passing mechanisms, restricting their architectures and expressive power. This work explores higher-rank irreducible Cartesian tensors to address these limitations. We integrate irreducible Cartesian tensor products into message-passing neural networks and prove the equivariance and traceless property of the resulting layers. Through empirical evaluations on various benchmark data sets, we consistently observe on-par or better performance than that of state-of-the-art spherical and Cartesian models.

cs.LG↗

arXiv2024

Uncertainty-biased molecular dynamics for learning uniformly accurate interatomic potentials

Efficiently creating a concise but comprehensive data set for training machine-learned interatomic potentials (MLIPs) is an under-explored problem. Active learning, which uses biased or unbiased molecular dynamics (MD) to generate candidate pools, aims to address this objective. Existing biased and unbiased MD-simulation methods, however, are prone to miss either rare events or extrapolative regions -- areas of the configurational space where unreliable predictions are made. This work demonstrates that MD, when biased by the MLIP's energy uncertainty, simultaneously captures extrapolative regions and rare events, which is crucial for developing uniformly accurate MLIPs. Furthermore, exploiting automatic differentiation, we enhance bias-forces-driven MD with the concept of bias stress. We employ calibrated gradient-based uncertainties to yield MLIPs with similar or, sometimes, better accuracy than ensemble-based methods at a lower computational cost. Finally, we apply uncertainty-biased MD to alanine dipeptide and MIL-53(Al), generating MLIPs that represent both configurational spaces more accurately than models trained with conventional MD.

physics.comp-ph↗

arXiv2024

PDEBENCH: An Extensive Benchmark for Scientific Machine Learning

Machine learning-based modeling of physical systems has experienced increased interest in recent years. Despite some impressive progress, there is still a lack of benchmarks for Scientific ML that are easy to use but still challenging and representative of a wide range of problems. We introduce PDEBench, a benchmark suite of time-dependent simulation tasks based on Partial Differential Equations (PDEs). PDEBench comprises both code and data to benchmark the performance of novel machine learning models against both classical numerical simulations and machine learning baselines. Our proposed set of benchmark problems contribute the following unique features: (1) A much wider range of PDEs compared to existing benchmarks, ranging from relatively common examples to more realistic and difficult problems; (2) much larger ready-to-use datasets compared to prior work, comprising multiple simulation runs across a larger number of initial and boundary conditions and PDE parameters; (3) more extensible source codes with user-friendly APIs for data generation and baseline results with popular machine learning models (FNO, U-Net, PINN, Gradient-Based Inverse Method). PDEBench allows researchers to extend the benchmark freely for their own purposes using a standardized API and to compare the performance of new models to existing baseline methods. We also propose new evaluation metrics with the aim to provide a more holistic understanding of learning methods in the context of Scientific ML. With those metrics we identify tasks which are challenging for recent ML methods and propose these tasks as future challenges for the community. The code is available at https://github.com/pdebench/PDEBench.

cs.LG↗

arXiv2023

Learning Neural PDE Solvers with Parameter-Guided Channel Attention

Scientific Machine Learning (SciML) is concerned with the development of learned emulators of physical systems governed by partial differential equations (PDE). In application domains such as weather forecasting, molecular dynamics, and inverse design, ML-based surrogate models are increasingly used to augment or replace inefficient and often non-differentiable numerical simulation algorithms. While a number of ML-based methods for approximating the solutions of PDEs have been proposed in recent years, they typically do not adapt to the parameters of the PDEs, making it difficult to generalize to PDE parameters not seen during training. We propose a Channel Attention mechanism guided by PDE Parameter Embeddings (CAPE) component for neural surrogate models and a simple yet effective curriculum learning strategy. The CAPE module can be combined with neural PDE solvers allowing them to adapt to unseen PDE parameters. The curriculum learning strategy provides a seamless transition between teacher-forcing and fully auto-regressive training. We compare CAPE in conjunction with the curriculum learning strategy using a popular PDE benchmark and obtain consistent and significant improvements over the baseline models. The experiments also show several advantages of CAPE, such as its increased ability to generalize to unseen PDE parameters without large increases inference time and parameter count.

cs.LG↗

arXiv2022

milIE: Modular & Iterative Multilingual Open Information Extraction

Open Information Extraction (OpenIE) is the task of extracting (subject, predicate, object) triples from natural language sentences. Current OpenIE systems extract all triple slots independently. In contrast, we explore the hypothesis that it may be beneficial to extract triple slots iteratively: first extract easy slots, followed by the difficult ones by conditioning on the easy slots, and therefore achieve a better overall extraction. Based on this hypothesis, we propose a neural OpenIE system, milIE, that operates in an iterative fashion. Due to the iterative nature, the system is also modular -- it is possible to seamlessly integrate rule based extraction systems with a neural end-to-end system, thereby allowing rule based systems to supply extraction slots which milIE can leverage for extracting the remaining slots. We confirm our hypothesis empirically: milIE outperforms SOTA systems on multiple languages ranging from Chinese to Arabic. Additionally, we are the first to provide an OpenIE test dataset for Arabic and Galician.

cs.CL↗

arXiv2021

An Empirical Study of the Effects of Sample-Mixing Methods for Efficient Training of Generative Adversarial Networks

It is well-known that training of generative adversarial networks (GANs) requires huge iterations before the generator's providing good-quality samples. Although there are several studies to tackle this problem, there is still no universal solution. In this paper, we investigated the effect of sample mixing methods, that is, Mixup, CutMix, and newly proposed Smoothed Regional Mix (SRMix), to alleviate this problem. The sample-mixing methods are known to enhance the accuracy and robustness in the wide range of classification problems, and can naturally be applicable to GANs because the role of the discriminator can be interpreted as the classification between real and fake samples. We also proposed a new formalism applying the sample-mixing methods to GANs with the saturated losses which do not have a clear "label" of real and fake. We performed a vast amount of numerical experiments using LSUN and CelebA datasets. The results showed that Mixup and SRMix improved the quality of the generated images in terms of FID in most cases, in particular, SRMix showed the best improvement in most cases. Our analysis indicates that the mixed-samples can provide different properties from the vanilla fake samples, and the mixing pattern strongly affects the decision of the discriminators. The generated images of Mixup have good high-level feature but low-level feature is not so impressible. On the other hand, CutMix showed the opposite tendency. Our SRMix showed the middle tendency, that is, showed good high and low level features. We believe that our finding provides a new perspective to accelerate the GANs convergence and improve the quality of generated samples.

cs.CV↗

arXiv2020

An Efficient Method of Training Small Models for Regression Problems with Knowledge Distillation

Compressing deep neural network (DNN) models becomes a very important and necessary technique for real-world applications, such as deploying those models on mobile devices. Knowledge distillation is one of the most popular methods for model compression, and many studies have been made on developing this technique. However, those studies mainly focused on classification problems, and very few attempts have been made on regression problems, although there are many application of DNNs on regression problems. In this paper, we propose a new formalism of knowledge distillation for regression problems. First, we propose a new loss function, teacher outlier rejection loss, which rejects outliers in training samples using teacher model predictions. Second, we consider a multi-task network with two outputs: one estimates training labels which is in general contaminated by noisy labels; And the other estimates teacher model's output which is expected to modify the noise labels following the memorization effects. By considering the multi-task network, training of the feature extraction of student models becomes more effective, and it allows us to obtain a better student model than one trained from scratch. We performed comprehensive evaluation with one simple toy model: sinusoidal function, and two open datasets: MPIIGaze, and Multi-PIE. Our results show consistent improvement in accuracy regardless of the annotation error level in the datasets.

cs.LG↗

arXiv2019

Evolution of three-dimensional Relativistic Ion Weibel Instability: Competition with Kink Instability

In this paper, we report our recent findings on the relativistic Weibel instability and its nonlinear saturation by performing numerical simulations of collisionless plasmas. Analysis of the obtained numerical results revealed that the nonlinear phase of the Weibel instability can be described by characteristic phases based on the Weibel filaments' current density in terms of particle and Alfvén limit currents. We also analyzed the relativistic kink instability based on the energy principle in the magnetohydrodynamic (MHD) regime, and found that the Weibel filaments do not suffer from the kink-type instability in the MHD regime up to 1000 $ω_{p,i}^{-1}$. This finding allowed a magnetic field to be sustained by relativistic Weibel instability that was stable enough to be a seed for MHD dynamos.

astro-ph.HE↗

arXiv2018

Magnetic Field Saturation of the Ion Weibel Instability in Interpenetrating Relativistic Plasmas

The time evolution and saturation of the Weibel instability at the ion Alfvén current are presented by ab initio particle-in-cell simulations. We found that the ion Weibel current in 3D could evolve into the Alfvén current where the magnetic field energy is sustained at 1.5\% of the initial beam kinetic energy. The current filaments are no longer isolated at saturation, but rather connected to each other to form a network structure. Electrons are continuously heated during the coalescence of the filaments, which is crucial for obtaining sustained magnetic fields with much stronger levels than with 2D simulations. The results highlight again the importance of the Weibel instability in generating magnetic fields in laboratory, astrophysical, and cosmological situations.

astro-ph.HE↗

arXiv2018

Evolution of 3-dimensional Relativistic Current Sheets and Development of Self-Generated Turbulence

In this paper, the temporal evolution of 3-dimensional relativistic current sheets in Poynting-dominated plasma is studied for the first time. Over the past few decades, a lot of efforts have been conducted on studying the evolution of current sheets in 2-dimensional space, and concluded that sufficiently long current sheets always evolves into the so-called "plasmoid-chain", which provides fast reconnection rate independent of its resistivity. However, it is suspected that plasmoid-chain can exist only in the case of 2-dimensional approximation, and would show transition to turbulence in 3-dimensional space. We performed 3-dimensional numerical simulation of relativistic current sheet using resistive relativistic magnetohydrodynamic approximation. The results showed that the 3-dimensional current sheet evolve not into plasmoid-chain but turbulence. The resulting reconnection rate is $0.004$ which is much smaller than that of plasmoid-chain. The energy conversion from magnetic field to kinetic energy of turbulence is just 0.01\% which is much smaller than typical non-relativistic cases. Using the energy principle, we also showed that the plasmoid is always unstable for a displacement in opposite direction to its acceleration, probably interchange-type instability, and this always results in seeds of turbulence behind the plasmoids. Finally, the temperature distribution along the sheet is discussed, and it is found that the sheet is less active than plasmoid-chain. Our finding can be applied for many high energy astrophysical phenomena, and can provide a basic model of the general current sheet in Poynting-dominated plasma.

astro-ph.HE↗

arXiv2017

Strong Coupling of Alfvén and Fast Modes in Compressible Relativistic Magnetohydrodynamic Turbulence in Magnetically-Dominated Plasmas

In this paper, we report our detailed analysis of the new strong-coupling regime between Alfvén and fast modes in Poynting-dominated plasma turbulence, reported in our previous work Takamoto & Lazarian (2016), which is an important effect for many relativistic plasma phenomena, and calls for new theories of Poynting-dominated MHD turbulence. We performed numerical simulations of relativistic MHD turbulence in isothermal plasmas, and analyzed the ratio of fast to Alfvén mode energy. We found that the increase of the fast mode with the background $σ$-parameter can be observed even in isothermal plasma, showing that such a phenomena is universal in trans-Alfvénic turbulence in Poynting-dominated plasmas. To study the detailed energy conversion process, we also performed a series of simulations of decaying turbulence injecting pure Alfvén, fast, and slow modes, respectively, and investigated the development of the mode conversion from the each mode. We also found that the mode conversion between Alfvén and fast modes is nearly insensitive to the background temperature. Finally, we report a result of a simulation with initially fast mode dominated turbulence. It developed into a temporally strong-coupling regime, which is a strong evidence for the existence of our suggesting strong-coupling regime of fast and Alfvén modes. Our result suggests that the strong turbulence in Poynting-dominated plasma is very different from that in the non-relativistic plasma. It will also give an important guidance to studies of particle acceleration and non-thermal photon emission from Poynting-dominated plasma.

astro-ph.HE↗

arXiv2016

Compressible Relativistic Magnetohydrodynamic Turbulence in Magnetically-Dominated Plasmas And Implications for A Strong-Coupling Regime

In this Letter, we report compressible mode effects on relativistic magnetohydrodynamic (RMHD) turbulence in Poynting-dominated plasmas using 3-dimensional numerical simulations. We decomposed fluctuations in the turbulence into 3 MHD modes (fast, slow, and Alfvén) following the procedure mode decomposition in (Cho & Lazarian 2002), and analyzed their energy spectra and structure functions separately. We also analyzed the ratio of compressible mode to Alfvén mode energy with respect to its Mach number. We found the ratio of compressible mode increases not only with the Alfvén Mach number but with the background magnetization, which indicates a strong coupling between the fast and Alfvén modes and appearance of a new regime of RMHD turbulence in Poynting-dominated plasmas where the fast and Alfvén modes strongly couples and cannot be separated, different from the non-relativistic MHD case. This finding will affect particle acceleration efficiency obtained by assuming Alfvénic critical balance turbulence, and will change the resulting photon spectra by non-thermal electrons.

astro-ph.HE↗

arXiv2015

Turbulent Reconnection in Relativistic Plasmas And Effects of Compressibility

We report turbulence effects on magnetic reconnection in relativistic plasmas using 3-dimensional relativistic resistive magnetohydrodynamics simulations. We found reconnection rate became independent of the plasma resistivity due to turbulence effects similarly to non-relativistic cases. We also found compressible turbulence effects modified the turbulent reconnection rate predicted in non-relativistic incompressible plasmas; The reconnection rate saturates and even decays as the injected velocity approaches to the Alfvén velocity. Our results indicate the compressibility cannot be neglected when compressible component becomes about half of incompressible mode occurring when the Alfvén Mach number reaches about $0.3$. The obtained maximum reconnection rate is around $0.05$ to $0.1$, which will be able to reach around $0.1$ to $0.2$ if injection scales are comparable to the sheet length.

astro-ph.HE↗