Search

Conventional medical artificial intelligence (AI) models face barriers in clinical application and ethical issues owing to their inability to handle the privacy-sensitive characteristics of medical data. We present a novel personalized federated learning (pFL) method for medical visual question answering (VQA) models, addressing privacy reliability challenges in the medical domain. Our method introduces learnable prompts into a Transformer architecture to efficiently train it on diverse medical datasets without massive computational costs. Then we introduce a reliable client VQA model that incorporates Dempster-Shafer evidence theory to quantify uncertainty in predictions, enhancing the model's reliability. Furthermore, we propose a novel inter-client communication mechanism that uses maximum likelihood estimation to balance accuracy and uncertainty, fostering efficient integration of insights across clients.

We propose a new strategy called think twice before recognizing to improve fine-grained traffic sign recognition (TSR). Fine-grained TSR in the wild is difficult due to the complex road conditions, and existing approaches particularly struggle with cross-country TSR when data is lacking. Our strategy achieves effective fine-grained TSR by stimulating the multiple-thinking capability of large multimodal models (LMM). We introduce context, characteristic, and differential descriptions to design multiple thinking processes for the LMM. The context descriptions with center coordinate prompt optimization help the LMM to locate the target traffic sign in the original road images containing multiple traffic signs and filter irrelevant answers through the proposed prior traffic sign hypothesis. The characteristic description is based on few-shot in-context learning of template traffic signs, which decreases the cross-domain difference and enhances the fine-grained recognition capability of the LMM. The differential descriptions of similar traffic signs optimize the multimodal thinking capability of the LMM. The proposed method is independent of training data and requires only simple and uniform instructions. We conducted extensive experiments on three benchmark datasets and two real-world datasets from different countries, and the proposed method achieves state-of-the-art TSR results on all five datasets.

We propose a novel symbolic music representation and Generative Adversarial Network (GAN) framework specially designed for symbolic multitrack music generation. The main theme of symbolic music generation primarily encompasses the preprocessing of music data and the implementation of a deep learning framework. Current techniques dedicated to symbolic music generation generally encounter two significant challenges: training data's lack of information about chords and scales and the requirement of specially designed model architecture adapted to the unique format of symbolic music representation. In this paper, we solve the above problems by introducing new symbolic music representation with MusicLang chord analysis model. We propose our MMT-BERT architecture adapting to the representation. To build a robust multitrack music generator, we fine-tune a pre-trained MusicBERT model to serve as the discriminator, and incorporate relativistic standard loss. This approach, supported by the in-depth understanding of symbolic music encoded within MusicBERT, fortifies the consonance and humanity of music generated by our method. Experimental results demonstrate the effectiveness of our approach which strictly follows the state-of-the-art methods., Comment: Accepted to the 25th International Society for Music Information Retrieval Conference (ISMIR 2024)

This paper presents our method for the generative track of The First Dataset Distillation Challenge at ECCV 2024. Since the diffusion model has become the mainstay of generative models because of its high-quality generative effects, we focus on distillation methods based on the diffusion model. Considering that the track can only generate a fixed number of images in 10 minutes using a generative model for CIFAR-100 and Tiny-ImageNet datasets, we need to use a generative model that can generate images at high speed. In this study, we proposed a novel generative dataset distillation method based on Stable Diffusion. Specifically, we use the SDXL-Turbo model which can generate images at high speed and quality. Compared to other diffusion models that can only generate images per class (IPC) = 1, our method can achieve an IPC = 10 for Tiny-ImageNet and an IPC = 20 for CIFAR-100, respectively. Additionally, to generate high-quality distilled datasets for CIFAR-100 and Tiny-ImageNet, we use the class information as text prompts and post data augmentation for the SDXL-Turbo model. Experimental results show the effectiveness of the proposed method, and we achieved third place in the generative track of the ECCV 2024 DD Challenge. Codes are available at https://github.com/Guang000/BANKO., Comment: The Third Place Winner in Generative Track of the ECCV 2024 DD Challenge

This paper reports on a measurement of electron-ion recombination in liquid argon in the ICARUS liquid argon time projection chamber (LArTPC). A clear dependence of recombination on the angle of the ionizing particle track relative to the drift electric field is observed. An ellipsoid modified box (EMB) model of recombination describes the data across all measured angles. These measurements are used for the calorimetric energy scale calibration of the ICARUS TPC, which is also presented. The impact of the EMB model is studied on calorimetric particle identification, as well as muon and proton energy measurements. Accounting for the angular dependence in EMB recombination improves the accuracy and precision of these measurements.

The ICARUS liquid argon time projection chamber (LArTPC) neutrino detector has been taking physics data since 2022 as part of the Short-Baseline Neutrino (SBN) Program. This paper details the equalization of the response to charge in the ICARUS time projection chamber (TPC), as well as data-driven tuning of the simulation of ionization charge signals and electronics noise. The equalization procedure removes non-uniformities in the ICARUS TPC response to charge in space and time. This work leverages the copious number of cosmic ray muons available to ICARUS at the surface. The ionization signal shape simulation applies a novel procedure that tunes the simulation to match what is measured in data. The end result of the equalization procedure and simulation tuning allows for a comparison of charge measurements in ICARUS between Monte Carlo simulation and data, showing good performance with minimal residual bias between the two.

Recent multimodal large language models (MLLM) such as GPT-4o and GPT-4v have shown great potential in autonomous driving. In this paper, we propose a cross-domain few-shot in-context learning method based on the MLLM for enhancing traffic sign recognition (TSR). We first construct a traffic sign detection network based on Vision Transformer Adapter and an extraction module to extract traffic signs from the original road images. To reduce the dependence on training data and improve the performance stability of cross-country TSR, we introduce a cross-domain few-shot in-context learning method based on the MLLM. To enhance MLLM's fine-grained recognition ability of traffic signs, the proposed method generates corresponding description texts using template traffic signs. These description texts contain key information about the shape, color, and composition of traffic signs, which can stimulate the ability of MLLM to perceive fine-grained traffic sign categories. By using the description texts, our method reduces the cross-domain differences between template and real traffic signs. Our approach requires only simple and uniform textual indications, without the need for large-scale traffic sign images and labels. We perform comprehensive evaluations on the German traffic sign recognition benchmark dataset, the Belgium traffic sign dataset, and two real-world datasets taken from Japan. The experimental results show that our method significantly enhances the TSR performance.

This paper proposes a novel zero-shot composed image retrieval (CIR) method considering the query-target relationship by masked image-text pairs. The objective of CIR is to retrieve the target image using a query image and a query text. Existing methods use a textual inversion network to convert the query image into a pseudo word to compose the image and text and use a pre-trained visual-language model to realize the retrieval. However, they do not consider the query-target relationship to train the textual inversion network to acquire information for retrieval. In this paper, we propose a novel zero-shot CIR method that is trained end-to-end using masked image-text pairs. By exploiting the abundant image-text pairs that are convenient to obtain with a masking strategy for learning the query-target relationship, it is expected that accurate zero-shot CIR using a retrieval-focused textual inversion network can be realized. Experimental results show the effectiveness of the proposed method., Comment: Accepted as a conference paper in IEEE ICIP 2024

This paper proposes a novel framework to reinforce classification models using language-guided generated counterfactual images. Deep learning classification models are often trained using datasets that mirror real-world scenarios. In this training process, because learning is based solely on correlations with labels, there is a risk that models may learn spurious relationships, such as an overreliance on features not central to the subject, like background elements in images. However, due to the black-box nature of the decision-making process in deep learning models, identifying and addressing these vulnerabilities has been particularly challenging. We introduce a novel framework for reinforcing the classification models, which consists of a two-stage process. First, we identify model weaknesses by testing the model using the counterfactual image dataset, which is generated by perturbed image captions. Subsequently, we employ the counterfactual images as an augmented dataset to fine-tune and reinforce the classification model. Through extensive experiments on several classification models across various datasets, we revealed that fine-tuning with a small set of counterfactual images effectively strengthens the model., Comment: 6 pages, 4 figures

This study presents a novel approach to Generative Class Incremental Learning (GCIL) by introducing the forgetting mechanism, aimed at dynamically managing class information for better adaptation to streaming data. GCIL is one of the hot topics in the field of computer vision, and this is considered one of the crucial tasks in society, specifically the continual learning of generative models. The ability to forget is a crucial brain function that facilitates continual learning by selectively discarding less relevant information for humans. However, in the field of machine learning models, the concept of intentionally forgetting has not been extensively investigated. In this study we aim to bridge this gap by incorporating the forgetting mechanisms into GCIL, thereby examining their impact on the models' ability to learn in continual learning. Through our experiments, we have found that integrating the forgetting mechanisms significantly enhances the models' performance in acquiring new knowledge, underscoring the positive role that strategic forgetting plays in the process of continual learning.

In this paper, we propose a new dataset distillation method that considers balancing global structure and local details when distilling the information from a large dataset into a generative model. Dataset distillation has been proposed to reduce the size of the required dataset when training models. The conventional dataset distillation methods face the problem of long redeployment time and poor cross-architecture performance. Moreover, previous methods focused too much on the high-level semantic attributes between the synthetic dataset and the original dataset while ignoring the local features such as texture and shape. Based on the above understanding, we propose a new method for distilling the original image dataset into a generative model. Our method involves using a conditional generative adversarial network to generate the distilled dataset. Subsequently, we ensure balancing global structure and local details in the distillation process, continuously optimizing the generator for more information-dense dataset generation., Comment: Accepted by the 1st CVPR Workshop on Dataset Distillation

Variable selection methods are required in practical statistical modeling, to identify and include only the most relevant predictors, and then improving model interpretability. Such variable selection methods are typically employed in regression models, for instance in this article for the Poisson Log Normal model (PLN, Chiquet et al., 2021). This model aim to explain multivariate count data using dependent variables, and its utility was demonstrating in scientific fields such as ecology and agronomy. In the case of the PLN model, most recent papers focus on sparse networks inference through combination of the likelihood with a L1 -penalty on the precision matrix. In this paper, we propose to rely on a recent penalization method (SIC, O'Neill and Burke, 2023), which consists in smoothly approximating the L0-penalty, and that avoids the calibration of a tuning parameter with a cross-validation procedure. Moreover, this work focuses on the coefficient matrix of the PLN model and establishes an inference procedure ensuring effective variable selection performance, so that the resulting fitted model explaining multivariate count data using only relevant explanatory variables. Our proposal involves implementing a procedure that integrates the SIC penalization algorithm (epsilon-telescoping) and the PLN model fitting algorithm (a variational EM algorithm). To support our proposal, we provide theoretical results and insights about the penalization method, and we perform simulation studies to assess the method, which is also applied on real datasets.

Estimating force constants for crystal structures is crucial for calculating various phonon-related properties. However, this task becomes particularly challenging when dealing with a large number of atoms or when third- and higher-order force constants are required. In this study, we propose an efficient approach that involves constructing a complete orthonormal basis set for the force constants. We formulate this approach using projection matrices and their eigenvectors to meet the requirements for the force constants. This basis set enables precise inference of the force constants from displacement-force datasets. Our efficient algorithms for basis-set construction and force constant estimation are implemented in the symfc code. Furthermore, several applications demonstrated in this study indicate that the current approach facilitates the efficient and accurate determination of force constants., Comment: 20 pages, 11 figures

We report on a search for magnetic monopoles (MMs) produced in ultraperipheral Pb--Pb collisions during Run-1 of the LHC. The beam pipe surrounding the interaction region of the CMS experiment was exposed to 184.07 \textmu b$^{-1}$ of Pb--Pb collisions at 2.76 TeV center-of-mass energy per collision in December 2011, before being removed in 2013. It was scanned by the MoEDAL experiment using a SQUID magnetometer to search for trapped MMs. No MM signal was observed. The two distinctive features of this search are the use of a trapping volume very close to the collision point and ultra-high magnetic fields generated during the heavy-ion run that could produce MMs via the Schwinger effect. These two advantages allowed setting the first reliable, world-leading mass limits on MMs with high magnetic charge. In particular, the established limits are the strongest available in the range between 2 and 45 Dirac units, excluding MMs with masses of up to 80 GeV at 95\% confidence level., Comment: As accepted by PRL. 7 pages, 4 figures

We present a novel prompt-based personalized federated learning (pFL) method to address data heterogeneity and privacy concerns in traditional medical visual question answering (VQA) methods. Specifically, we regard medical datasets from different organs as clients and use pFL to train personalized transformer-based VQA models for each client. To address the high computational complexity of client-to-client communication in previous pFL methods, we propose a succinct information sharing system by introducing prompts that are small learnable parameters. In addition, the proposed method introduces a reliability parameter to prevent the negative effects of low performance and irrelevant clients. Finally, extensive evaluations on various heterogeneous medical datasets attest to the effectiveness of our proposed method., Comment: Accept by ICASSP2024

The application of first-principles calculations for predicting lattice thermal conductivity (LTC) in crystalline materials, in conjunction with the linearized phonon Boltzmann equation, has gained increasing popularity. In this calculation, the determination of force constants through first-principles calculations is critical for accurate LTC predictions. For material exploration, performing first-principles LTC calculations in a high-throughput manner is now expected, although it requires significant computational resources. To reduce computational demands, we integrated polynomial machine learning potentials on-the-fly during the first-principles LTC calculations. This paper presents a systematic approach to first-principles LTC calculations. We designed and optimized an efficient workflow that integrates multiple modular software packages. We applied this approach to calculate LTCs for 103 compounds of the wurtzite, zincblende, and rocksalt types to evaluate the performance of the polynomial machine learning potentials in LTC calculations. We demonstrate a significant reduction in the computational resources required for the LTC predictions.

In this paper, the time-dependent dielectric breakdown (TDDB) in sub-1-nm equivalent oxide thickness (EOT) n-type bulk FinFETs is studied. The gate stacks consist of an IMEC clean interfacial layer, atomic layer deposition HfO2 high-k and TiN metal electrode. For the 0.8-nm EOT FinFETs, it is found that TDDB lifetime is consistent with results of planar devices for areas around 1e-8 cm2, implying that the FinFET architecture does not seem to introduce new failure mechanisms. However, for devices with smaller area, the extrapolated voltage at a ten-year lifetime for soft breakdown (SBD) does not meet the specifications, and as a consequence, the SBD path wear-out will have to be included in the final extrapolation. Furthermore, it is shown that for EOTs smaller than 0.8 nm, the TDDB reliability on n-type FinFETs is challenged by the high leakage currents., Comment: 15 pages, 7 figures

The symmetry of a crystal structure with a three-dimensional (3D) lattice can be classified by one of the 230 space group types. For some types of crystals, e.g. crystalline films, surfaces, or planar interfaces, it is more appropriate to assume a 2D lattice. With this assumption, the structure can be classified by one of the 80 layer group types. We have implemented an algorithm to determine the layer group type of a 3D structure with a 2D lattice, and applied it to more than 15,000 monolayer structures in the Computational 2D Materials (C2DB) database. We compare the classification of monolayers by layer groups and space groups, respectively. The latter is defined as the space group of the 3D bulk structure obtained by repeating the monolayer periodically in the direction perpendicular to the 2D lattice (AA-stacking). By this correspondence, nine pairs of layer group types are mapped to the same space group type due to the inability of the space group to distinguish the in-plane and out-of-plane axes. In total 18% of the monolayers in the C2DB belong to one of these layer group pairs and are thus not properly classified by the space group type.

Herein, we propose a novel dataset distillation method for constructing small informative datasets that preserve the information of the large original datasets. The development of deep learning models is enabled by the availability of large-scale datasets. Despite unprecedented success, large-scale datasets considerably increase the storage and transmission costs, resulting in a cumbersome model training process. Moreover, using raw data for training raises privacy and copyright concerns. To address these issues, a new task named dataset distillation has been introduced, aiming to synthesize a compact dataset that retains the essential information from the large original dataset. State-of-the-art (SOTA) dataset distillation methods have been proposed by matching gradients or network parameters obtained during training on real and synthetic datasets. The contribution of different network parameters to the distillation process varies, and uniformly treating them leads to degraded distillation performance. Based on this observation, we propose an importance-aware adaptive dataset distillation (IADD) method that can improve distillation performance by automatically assigning importance weights to different network parameters during distillation, thereby synthesizing more robust distilled datasets. IADD demonstrates superior performance over other SOTA dataset distillation methods based on parameter matching on multiple benchmark datasets and outperforms them in terms of cross-architecture generalization. In addition, the analysis of self-adaptive weights demonstrates the effectiveness of IADD. Furthermore, the effectiveness of IADD is validated in a real-world medical application such as COVID-19 detection., Comment: Published as a journal paper in Elsevier Neural Networks

A spin space group provides a suitable way to fully exploit the symmetry of a spin arrangement with a negligible spin-orbit coupling. There has been a growing interest in applying spin symmetry analysis with the spin space group in the field of magnetism. However, there is no established algorithm to search for spin symmetry operations of the spin space group. This paper presents an exhaustive algorithm for determining spin symmetry operations of commensurate spin arrangements. The present algorithm searches for spin symmetry operations from the symmetry operations of a corresponding nonmagnetic crystal structure and determines their spin-rotation parts by solving a Procrustes problem. An implementation is distributed under a permissive free software license in spinspg v0.1.1: https://github.com/spglib/spinspg.

Recent studies identified spin-order-driven phenomena such as spin-charge interconversion without relying on the relativistic spin-orbit interaction. Those physical properties can be prominent in systems containing light magnetic atoms due to sizable exchange splitting and may pave the way for realization of giant responses correlated with the spin degree of freedom. In this paper, we present a systematic symmetry analysis based on the spin crystallographic groups and identify the physical property of a vast number of magnetic materials up to 1500 in total. By decoupling the spin and orbital degrees of freedom, our analysis enables us to take a closer look into the relation between the dimensionality of spin structures and the resultant physical properties and to identify the spin and orbital contributions separately. In stark contrast to the established analysis with magnetic space groups, the spin crystallographic group manifests richer symmetry including spin-translation symmetry and leads to emergent responses. For representative examples, we discuss the geometrical nature of the anomalous Hall effect and magnetoelectric effect and classify the spin Hall effect arising from the nonrelativistic spin-charge coupling. Using the power of computational analysis, we apply our symmetry analysis to a wide range of magnets, encompassing complex magnets such as those with noncoplanar spin structures as well as collinear and coplanar magnets. We identify emergent multipoles relevant to physical responses and argue that our method provides a systematic tool for exploring sizable electromagnetic responses driven by spin order., Comment: 33 pages, 7 figures, 6 tables