Search

The proliferation of Text-to-Image (T2I) models has revolutionized content creation, providing powerful tools for diverse applications ranging from artistic expression to educational material development and marketing. Despite these technological advancements, significant ethical concerns arise from these models' reliance on large-scale datasets that often contain inherent societal biases. These biases are further amplified when AI-generated content is included in training data, potentially reinforcing and perpetuating stereotypes in the generated outputs. In this paper, we introduce FairT2I, a novel framework that harnesses large language models to detect and mitigate social biases in T2I generation. Our framework comprises two key components: (1) an LLM-based bias detection module that identifies potential social biases in generated images based on text prompts, and (2) an attribute rebalancing module that fine-tunes sensitive attributes within the T2I model to mitigate identified biases. Our extensive experiments across various T2I models and datasets show that FairT2I can significantly reduce bias while maintaining high-quality image generation. We conducted both qualitative user studies and quantitative non-parametric analyses in the generated image feature space, building upon the occupational dataset introduced in the Stable Bias study. Our results show that FairT2I successfully mitigates social biases and enhances the diversity of sensitive attributes in generated images. We further demonstrate, using the P2 dataset, that our framework can detect subtle biases that are challenging for human observers to perceive, extending beyond occupation-related prompts. On the basis of these findings, we introduce a new benchmark dataset for evaluating bias in T2I models.

We propose a novel mechanism of primordial black hole (PBH) formation through inverted bubble collapse. In this scenario, bubbles nucleate sparsely in an incomplete first-order phase transition, followed by a bulk phase transition in the rest of the universe that inverts these pre-existing bubbles into false vacuum regions. These spherically symmetric false-vacuum bubbles subsequently collapse to form PBHs. Unlike conventional PBH formation mechanisms associated with domain wall collapse or bubble coalescence, our inverted bubble collapse mechanism naturally ensures spherical collapse. We demonstrate that, when applied to the electroweak phase transition, this mechanism can produce highly monochromatic PBHs with masses up to ${\cal O}(10^{-6}\,\text{-}\,10^{-5}) M_\odot$, which potentially explain the microlensing events observed in the OGLE and Subaru HSC data., Comment: 8 pages, 3 figures

The camera of the Large-Sized Telescopes (LSTs) of the Cherenkov Telescope Array Observatory (CTAO) consists of 1855 pixels that are grouped into 265 high-performance photomultiplier tube (PMT) modules. Each module comprises a seven-light-guide plate, seven PMT units, a slow control board, and a readout board with a trigger board. %In this paper we describe The requirements for the PMT modules include various aspects, such as photon detection efficiency, dynamic range, buffer depth, and test pulse functionality. We have developed a high-performance PMT module that fulfills all these requirements. Mass-production and quality control (QC) of modules for all four LSTs of the northern CTAO have been completed. Here we report on the technical details of each element of the module and its performance, together with the methods and results of QC measurements., Comment: Published in NIM A

Photomultiplier tubes (PMTs) have been widely used in imaging atmospheric Cherenkov telescopes (IACTs). The Large-Sized Telescopes (LSTs) of the Cherenkov Telescope Array Observatory (CTAO), the latest-generation IACTs, are optimized for challenging observations of low-energy gamma rays, specifically in the 20 to 150 GeV range. To this end, PMTs with an exceptionally low afterpulsing probability have been developed and installed. However, the afterpulsing rate increases over time due to the infiltration of atmospheric molecules, particularly helium, into the tube. Interestingly, we found that the afterpulsing rate decreases when PMTs are operated at high voltage and exposed to light -- a condition naturally met during IACT observations. To evaluate the latest instrument response, after five years of operation, we removed several PMTs from the first LST, which is currently the only operational telescope among the CTAO instruments. Our laboratory measurements showed no increase in afterpulsing compared to pre-installation values. This suggests that the decrease in afterpulsing during operation offsets the increase, thereby maintaining the long-term performance of the PMTs., Comment: The 6th International Workshop on New Photon-Detector (PD24)

Segment Anything Model (SAM) represents a large-scale segmentation model that enables powerful zero-shot capabilities with flexible prompts. While SAM can segment any object in zero-shot, it requires user-provided prompts for each target image and does not attach any label information to masks. Few-shot segmentation models addressed these issues by inputting annotated reference images as prompts to SAM and can segment specific objects in target images without user-provided prompts. Previous SAM-based few-shot segmentation models only use annotated reference images as prompts, resulting in limited accuracy due to a lack of reference information. In this paper, we propose a novel few-shot segmentation model, Vision and Language reference Prompt into SAM (VLP-SAM), that utilizes the visual information of the reference images and the semantic information of the text labels by inputting not only images but also language as reference information. In particular, VLP-SAM is a simple and scalable structure with minimal learnable parameters, which inputs prompt embeddings with vision-language information into SAM using a multimodal vision-language model. To demonstrate the effectiveness of VLP-SAM, we conducted experiments on the PASCAL-5i and COCO-20i datasets, and achieved high performance in the few-shot segmentation task, outperforming the previous state-of-the-art model by a large margin (6.3% and 9.5% in mIoU, respectively). Furthermore, VLP-SAM demonstrates its generality in unseen objects that are not included in the training data. Our code is available at https://github.com/kosukesakurai1/VLP-SAM., Comment: 8 pages, 2 figures

The electric dipole moments~(EDM) of fundamental particles inherently violate parity~(P) and time-reversal~(T) symmetries. By virtue of the CPT theorem in quantum field theory, the latter also implies the violation of the combined charge-conjugation and parity~(CP) symmetry. We aim to measure the EDM of the muon using the frozen-spin technique within a compact storage trap. This method exploits the high effective electric field, \$E \approx 165\$ MV/m, experienced in the rest frame of the muon with a momentum of about 23 MeV/c when it passes through a solenoidal magnetic field of \$|\vec{B}|=2.5\$ T. In this paper, we outline the fundamental considerations for a muon EDM search and present a conceptual design for a demonstration experiment to be conducted at secondary muon beamlines of the Paul Scherrer Institute in Switzerland. In Phase~I, with an anticipated data acquisition period of 200 days, the expected sensitivity to a muon EDM is 4E-21 ecm. In a subsequent phase, Phase~II, we propose to improve the sensitivity to 6E-23 ecm using a dedicated instrument installed on a different beamline that produces muons of momentum 125 MeV/c}., Comment: 34 pages, submitted to EPJC

Time-sensitive machine learning benefits from Sequential Probability Ratio Test (SPRT), which provides an optimal stopping time for early classification of time series. However, in finite horizon scenarios, where input lengths are finite, determining the optimal stopping rule becomes computationally intensive due to the need for backward induction, limiting practical applicability. We thus introduce FIRMBOUND, an SPRT-based framework that efficiently estimates the solution to backward induction from training data, bridging the gap between optimal stopping theory and real-world deployment. It employs density ratio estimation and convex function learning to provide statistically consistent estimators for sufficient statistic and conditional expectation, both essential for solving backward induction; consequently, FIRMBOUND minimizes Bayes risk to reach optimality. Additionally, we present a faster alternative using Gaussian process regression, which significantly reduces training time while retaining low deployment overhead, albeit with potential compromise in statistical consistency. Experiments across independent and identically distributed (i.i.d.), non-i.i.d., binary, multiclass, synthetic, and real-world datasets show that FIRMBOUND achieves optimalities in the sense of Bayes risk and speed-accuracy tradeoff. Furthermore, it advances the tradeoff boundary toward optimality when possible and reduces decision-time variance, ensuring reliable decision-making. Code is publicly available at https://github.com/Akinori-F-Ebihara/FIRMBOUND, Comment: Accepted to International Conference on Learning Representations (ICLR) 2025

Introduced in arXiv:2211.12606, biquandle arrow weight invariants are enhancements of the biquandle counting invariant for oriented virtual and classical knots defined from biquandle-colored Gauss diagrams using a tensor over an abelian group satisfying certain properties. In this paper we categorify the biquandle arrow weight polynomial invariant using biquandle coloring quivers, obtaining new infinite families of polynomial invariants of oriented virtual and classical knots., Comment: 11 pages; sequel to arXiv:2211.12606

The security of blockchain systems based on Proof of Work relies on mining. However, mining suffers from unstable revenue, prompting many miners to form cooperative mining pools. Most existing mining pools operate in a centralized manner, which undermines the decentralization principle of blockchain. Distributed mining pools offer a practical solution to this problem. Well-known examples include P2Pool and SmartPool. However, P2Pool encounters scalability and security issues in its early stages. Similarly, SmartPool is not budget-balanced and imposes fees due to its heavy use of the smart contract. In this research, we present a distributed mining pool named FiberPool to address these challenges. FiberPool integrates a smart contract on the main chain, a storage chain for sharing data necessary for share verification, and a child chain to reduce fees associated with using and withdrawing block rewards. We validate the mining fairness, budget balance, reward stability, and incentive compatibility of the payment scheme FiberPool Proportional adopted by FiberPool.

Anomaly detection is crucial in financial auditing and effective detection often requires obtaining large volumes of data from multiple organizations. However, confidentiality concerns hinder data sharing among audit firms. Although the federated learning (FL)-based approach, FedAvg, has been proposed to address this challenge, its use of mutiple communication rounds increases its overhead, limiting its practicality. In this study, we propose a novel framework employing Data Collaboration (DC) analysis -- a non-model share-type FL method -- to streamline model training into a single communication round. Our method first encodes journal entry data via dimensionality reduction to obtain secure intermediate representations, then transforms them into collaboration representations for building an autoencoder that detects anomalies. We evaluate our approach on a synthetic dataset and real journal entry data from multiple organizations. The results show that our method not only outperforms single-organization baselines but also exceeds FedAvg in non-i.i.d. experiments on real journal entry data that closely mirror real-world conditions. By preserving data confidentiality and reducing iterative communication, this study addresses a key auditing challenge -- ensuring data confidentiality while integrating knowledge from multiple audit firms. Our findings represent a significant advance in artificial intelligence-driven auditing and underscore the potential of FL methods in high-security domains.

We present numerical simulations of x-ray magnetic circular dichroism (XMCD) at the L$_{2,3}$ edge of Ni in the weakly ferromagnetic altermagnet NiF$_2$. Our results predict a significant XMCD signal for light propagating perpendicular to the magnetic moments, which are approximately aligned along the [100] easy-axis direction. The analysis shows that the altermagnetic and ferromagnetic contributions to the XMCD signal can be uniquely distinguished by their dependence on an applied magnetic field. By varying the angle of the field relative to the easy axis, the in-plane orientation of both the N\'eel vector and the net magnetization can be systematically controlled. We further demonstrate that the XMCD signal, even under fields as strong as 40 T and for any in-plane orientation, can be accurately described as a linear combination of two spectral components, with geometrical prefactors determined by the field magnitude and direction. This insight enables experimental validation of the distinctive relationship between the N\'eel vector orientation and the x-ray Hall vector in the rutile structure. Quantitative simulations supporting these findings are provided., Comment: 6 pages, 6 figures, supplemental material

Observational studies enable causal inferences when randomized controlled trials (RCTs) are not feasible. However, integrating sensitive medical data across multiple institutions introduces significant privacy challenges. The data collaboration quasi-experiment (DC-QE) framework addresses these concerns by sharing "intermediate representations" -- dimensionality-reduced data derived from raw data -- instead of the raw data. While the DC-QE can estimate treatment effects, its application to medical data remains unexplored. This study applied the DC-QE framework to medical data from a single institution to simulate distributed data environments under independent and identically distributed (IID) and non-IID conditions. We propose a novel method for generating intermediate representations within the DC-QE framework. Experimental results demonstrated that DC-QE consistently outperformed individual analyses across various accuracy metrics, closely approximating the performance of centralized analysis. The proposed method further improved performance, particularly under non-IID conditions. These outcomes highlight the potential of the DC-QE framework as a robust approach for privacy-preserving causal inferences in healthcare. Broader adoption of this framework and increased use of intermediate representations could grant researchers access to larger, more diverse datasets while safeguarding patient confidentiality. This approach may ultimately aid in identifying previously unrecognized causal relationships, support drug repurposing efforts, and enhance therapeutic interventions for rare diseases.

This paper demonstrates real-time short-term traffic flow prediction through distributed fiber-optic sensing (DFOS) and data assimilation with a stochastic cell-automata-based traffic model. Traffic congestion on expressways is a severe issue. To alleviate its negative impacts, it is necessary to optimize traffic flow prior to becoming serious congestion. For this purpose, real-time short-term traffic flow prediction is promising. However, conventional traffic monitoring apparatus used in prediction methods faces a technical issue due to the sparsity in traffic flow data. To overcome the issue for realizing real-time traffic prediction, this paper employs DFOS, which enables to obtain spatially continuous and real-time traffic flow data along the road without dead zones. Using mean velocities derived from DFOS data as a feature extraction, this paper proposes a real-time data assimilation method for the short-term prediction. As the theoretical model, the stochastic Nishinari-Fukui-Schadschneider model is adopted. Future traffic flow is simulated with the optimal values of model parameters estimated from observed mean velocities and the initial condition estimated as the latest microscopic traffic state. This concept is validated using two congestion scenarios obtained in Japanese expressways. The results show that the mean absolute error of the predicted mean velocities is 10-15 km/h in the prediction horizon of 30 minutes. Furthermore, the prediction error in congestion length and travel time decreases by 40-84% depending on congestion scenarios when compared with conventional methods with traffic counters. This paper concludes that real-time data assimilation using DFOS enables an accurate short-term traffic prediction., Comment: 22 pages, 11 figures

The connectivity between qubits plays a crucial role in the performance of quantum extreme reservoir computing (QERC), particularly regarding long-range and inter-modular connections. We demonstrate that sufficiently long-range connections within a single module can achieve performance comparable to fully connected networks in supervised learning tasks. Further analysis of inter-modular connection schemes -- such as boundary, parallel, and arbitrary links -- shows that even a small number of well-placed connections can significantly enhance QERC performance. These findings suggest that modular QERC architectures, which could be more easily implemented on two-dimensional quantum chips or through the integration of small quantum systems, provide an effective approach for machine learning tasks.

This study aims to improve the efficiency and quality of career interviews conducted by nursing managers. To this end, we have been developing a slot-filling dialogue system that engages in pre-interviews to collect information on staff careers as a preparatory step before the actual interviews. Conventional slot-filling-based interview dialogue systems have limitations in the flexibility of information collection because the dialogue progresses based on predefined slot sets. We therefore propose a method that leverages large language models (LLMs) to dynamically generate new slots according to the flow of the dialogue, achieving more natural conversations. Furthermore, we incorporate abduction into the slot generation process to enable more appropriate and effective slot generation. To validate the effectiveness of the proposed method, we conducted experiments using a user simulator. The results suggest that the proposed method using abduction is effective in enhancing both information-collecting capabilities and the naturalness of the dialogue., Comment: 9 pages, 9 tables, 2 figures; 14 pages of appendix. Accepted to COLING 2025

Deep learning models have recently shown great success in classifying epileptic patients using EEG recordings. Unfortunately, classification-based methods lack a sound mechanism to detect the onset of seizure events. In this work, we propose a two-stage framework, \method, that explicitly models seizure onset through a novel task formulation of subsequence clustering. Given an EEG sequence, the framework first learns a set of second-level embeddings with label supervision. It then employs model-based clustering to explicitly capture long-term temporal dependencies in EEG sequences and identify meaningful subsequences. Epochs within a subsequence share a common cluster assignment (normal or seizure), with cluster or state transitions representing successful onset detections. Extensive experiments on three datasets demonstrate that our method can correct misclassifications, achieving 5%-11% classification improvements over other baselines and accurately detecting seizure onsets., Comment: Accepted at AAAI 2025

Knowledge Graphs (KGs) represent relationships between entities in a graph structure and have been widely studied as promising tools for realizing recommendations that consider the accurate content information of items. However, traditional KG-based recommendation methods face fundamental challenges: insufficient consideration of temporal information and poor performance in cold-start scenarios. On the other hand, Large Language Models (LLMs) can be considered databases with a wealth of knowledge learned from the web data, and they have recently gained attention due to their potential application as recommendation systems. Although approaches that treat LLMs as recommendation systems can leverage LLMs' high recommendation literacy, their input token limitations make it impractical to consider the entire recommendation domain dataset and result in scalability issues. To address these challenges, we propose a LLM's Intuition-aware Knowledge graph Reasoning model (LIKR). Our main idea is to treat LLMs as reasoners that output intuitive exploration strategies for KGs. To integrate the knowledge of LLMs and KGs, we trained a recommendation agent through reinforcement learning using a reward function that integrates different recommendation strategies, including LLM's intuition and KG embeddings. By incorporating temporal awareness through prompt engineering and generating textual representations of user preferences from limited interactions, LIKR can improve recommendation performance in cold-start scenarios. Furthermore, LIKR can avoid scalability issues by using KGs to represent recommendation domain datasets and limiting the LLM's output to KG exploration strategies. Experiments on real-world datasets demonstrate that our model outperforms state-of-the-art recommendation methods in cold-start sequential recommendation scenarios., Comment: Accepted to the 47th European Conference on Information Retrieval (ECIR2025)

Mechanism design, a branch of economics, aims to design rules that can autonomously achieve desired outcomes in resource allocation and public decision making. The research on mechanism design using machine learning is called automated mechanism design or mechanism learning. In our research, we constructed a new network based on the existing method for single auctions and aimed to automatically design a mechanism by applying it to double auctions. In particular, we focused on the following four desirable properties for the mechanism: individual rationality, balanced budget, Pareto efficiency, and incentive compatibility. We conducted experiments assuming a small-scale double auction and clarified how deterministic the trade matching of the obtained mechanism is. We also confirmed how much the learnt mechanism satisfies the four properties compared to two representative protocols. As a result, we verified that the mechanism is more budget-balanced than the VCG protocol and more economically efficient than the MD protocol, with the incentive compatibility mostly guaranteed.

A novel oxychloride, Sr$_{3}$Ni$_{2}$O$_{5}$Cl$_{2}$, was synthesized for the first time under high pressure of 10 GPa at 1400 ${}^\circ$C, motivated by a theoretical prediction of its potential superconductivity under ambient pressure. Small single crystals were used to determine the crystal structure and measure the temperature dependence of electrical resistance. The crystal is isostructural with the recently discovered superconductor, La$_{3}$Ni$_{2}$O$_{7}$, in line with the theoretical expectation.

We study the problem of modeling a non-linear dynamical system when given a time series by deriving equations directly from the data. Despite the fact that time series data are given as input, models for dynamics and estimation algorithms that incorporate long-term temporal dependencies are largely absent from existing studies. In this paper, we introduce a latent state to allow time-dependent modeling and formulate this problem as a dynamics estimation problem in latent states. We face multiple technical challenges, including (1) modeling latent non-linear dynamics and (2) solving circular dependencies caused by the presence of latent states. To tackle these challenging problems, we propose a new method, Latent Non-Linear equation modeling (LaNoLem), that can model a latent non-linear dynamical system and a novel alternating minimization algorithm for effectively estimating latent states and model parameters. In addition, we introduce criteria to control model complexity without human intervention. Compared with the state-of-the-art model, LaNoLem achieves competitive performance for estimating dynamics while outperforming other methods in prediction., Comment: Accepted by AAAI'25

The first spectroscopy of $^{52}$K was investigated via in-beam $\gamma$-ray spectroscopy at the RIKEN Radioactive Isotope Beam Factory after one-proton and one-neutron knockout from $^{53}$Ca and $^{53}$K beams impinging on a 15-cm liquid hydrogen target at $\approx$ 230~MeV/nucleon. The energy level scheme of $^{52}$K was built using single $\gamma$ and $\gamma$-$\gamma$ coincidence spectra. The spins and parities of the excited states were established based on momentum distributions of the fragment after the knockout reaction and based on exclusive cross sections. The results were compared to state-of-the-art shell model calculations with the SDPF-Umod interaction and ab initio IMSRG calculations with chiral effective field theory nucleon-nucleon and three-nucleon forces.

In quantum information theory, the evolution of an open quantum system -- a unitary evolution followed by a measurement -- is described by a quantum channel or, more generally, a quantum instrument. In this work, we formulate spin and flavour measurements in collider experiments as a quantum instrument. We demonstrate that the Choi matrix, which completely determines input-output transitions, can be both theoretically computed from a given model and experimentally reconstructed from a set of final state measurements (quantum state tomography) using varied input states. The reconstruction of the Choi matrix, known as quantum process tomography, offers a powerful new approach for probing potential extensions of the Standard Model within the quantum field theory framework and also provides a fundamental test of quantum mechanics itself. As an example, we outline a quantum process tomography approach applied to the $e^+ e^- \to t \bar{t}$ process at a polarized lepton collider., Comment: 33 pages, 1 figure

Vibrational sum frequency generation (SFG) is a nonlinear spectroscopic technique widely used to study the molecular structure and dynamics of surface systems. However, the spatial resolution achieved by far-field observations is constrained by the diffraction limit, obscuring molecular details in inhomogeneous structures smaller than the wavelength of light. To overcome this limitation, we developed a system for tip-enhanced SFG (TE-SFG) spectroscopy based on a scanning tunneling microscope. We successfully detected vibrational TE-SFG signals from adsorbed molecules on a gold substrate under ambient conditions. The phase analysis of interferometric SFG spectra provided information on molecular orientation. Furthermore, the observed TE-SFG signal was confirmed to originate from a highly localized region within a gap between the tip apex and the sample substrate. This method offers a novel platform for nonlinear optical nanospectroscopy, paving the way for the investigation of surface molecular systems beyond the diffraction limit., Comment: The manuscript includes Supporting Information detailing the experimental procedures and data analysis

Recently, an increasing number of laws have governed the useability of users' privacy. For example, Article 17 of the General Data Protection Regulation (GDPR), the right to be forgotten, requires machine learning applications to remove a portion of data from a dataset and retrain it if the user makes such a request. Furthermore, from the security perspective, training data for machine learning models, i.e., data that may contain user privacy, should be effectively protected, including appropriate erasure. Therefore, researchers propose various privacy-preserving methods to deal with such issues as machine unlearning. This paper provides an in-depth review of the security and privacy concerns in machine learning models. First, we present how machine learning can use users' private data in daily life and the role that the GDPR plays in this problem. Then, we introduce the concept of machine unlearning by describing the security threats in machine learning models and how to protect users' privacy from being violated using machine learning platforms. As the core content of the paper, we introduce and analyze current machine unlearning approaches and several representative research results and discuss them in the context of the data lineage. Furthermore, we also discuss the future research challenges in this field.

An effective method for suppressing intentional forks in a blockchain is the last-generated rule, which selects the most recent chain as the main chain in the event of a chain tie. This rule helps invalidate blocks that are withheld by adversaries for a certain period. However, existing last-generated rules face an issue in that their applications to the system are not fully localized. In conservative cryptocurrency systems such as Bitcoin, it is desirable for methods to be applied in a fully local manner. In this paper, we propose a locally applicable last-generated rule. Our method is straightforward and is based on a relative time reference. By conservatively setting the upper bound for the clock skews $\Delta_{O_i}$ to 200 s, our proposed method reduces the proportion $\gamma$ of honest miners following the attacker during chain ties by more than 40% compared to existing local methods.

We discuss the thermal production of axions in renormalizable models involving two Higgs doublet fields and a complex singlet field with a global $U(1)$ Peccei-Quinn symmetry, i.e., DFSZ type axion models. We demonstrate that, when the reheating temperature exceeds the mass scale of heavy Higgs bosons, axions are efficiently produced through heavy Higgs boson decays and scatterings at temperatures comparable to the heavy Higgs boson mass scale. As a result, the abundance of thermally produced axions is independent of the reheating temperature, which should be contrasted with the KSVZ axion model. This is because thermal productions via renormalizable interactions are IR-dominated processes. We demonstrate that the heavy Higgs boson decays are the main channels for axion thermal productions among various processes in the DFSZ-type axion models, which were missed in the literature. Our results apply to the original DFSZ QCD axion model since the production mechanism does not depend on the axion mass. As an application of axion productions from the heavy Higgs boson decays, we calculate the contributions to $\Delta N_{\rm eff}$ for axions with a mass smaller than ${\cal O}(0.1){\rm eV}$. Future measurements of $\Delta N_{\rm eff}$ could constrain model parameters in both axion and Higgs sectors. Focusing on axions with masses from keV to sub-GeV scale, we then discuss how cosmological observations such as X-ray and cosmic microwave background constrain the produced axion. We show that a large portion of the parameter space of the models can be explored even if the amount of the axion produced from the heavy Higgs bosons is much smaller than the observed cold dark matter abundance., Comment: 44 pages, 10 figures, 1 table

Telekinesis is the ability to manipulate remote objects without direct physical contact. In fictional works, telekinesis users are often depicted as controlling objects with their hands and other body parts as if by will alone. Such depictions suggest that users experience a sense of agency over the object despite not physically touching it. In this study, we developed a VR method to simulate telekinesis and investigated whether it is possible to achieve a sense of physical sensation and agency similar to the experience portrayed in fiction., Comment: Part of proceedings of 6th International Conference AsiaHaptics 2024

This study presents a novel contact detection system for "Parablade," a chambara-based, sword-play augmented sport. Augmented sports combine physical activities with virtual parameters (VPs) to create a balanced and equitable gaming experience, irrespective of players' physical capabilities. The proposed Parablade Microphone Unit (PMU) employs multiple micro-phones and machine learning algorithms to detect and classify hit events through sound recogni-tion. This system aims to ensure real-time updates of VPs, thereby enhancing the gameplay expe-rience. Experimental results indicate that the PMU can accurately recognize the occurrence and location of hit events with a high accuracy rate of 93.33%, with the assistance of 10kHz additional sound generated from the sword., Comment: Part of proceedings of 6th International Conference AsiaHaptics 2024

While recent years have witnessed the advancement in big data and Artificial Intelligence (AI), it is of much importance to safeguard data privacy and security. As an innovative approach, Federated Learning (FL) addresses these concerns by facilitating collaborative model training across distributed data sources without transferring raw data. However, the challenges of robust security and privacy across decentralized networks catch significant attention in dealing with the distributed data in FL. In this paper, we conduct an extensive survey of the security and privacy issues prevalent in FL, underscoring the vulnerability of communication links and the potential for cyber threats. We delve into various defensive strategies to mitigate these risks, explore the applications of FL across different sectors, and propose research directions. We identify the intricate security challenges that arise within the FL frameworks, aiming to contribute to the development of secure and efficient FL systems., Comment: 32 pages, to appear in KAIS

Isospin symmetry is a fundamental concept in nuclear physics. Even though isospin symmetry is partially broken, it holds approximately for most nuclear systems, which makes exceptions very interesting from the nuclear structure perspective. In this framework, it is expected that the spins and parities of the ground states of mirror nuclei should be the same, in particular for the simplest systems where a proton is exchanged with a neutron or vice versa. In this work, we present evidence that this assumption is broken in the mirror pair $^{71}$Br and $^{71}$Kr system. Our conclusions are based on a high-statistics $\beta$ decay study of $^{71}$Kr and on state-of-the-art shell model calculations. In our work, we also found evidence of a new state in $^{70}$Se, populated in the $\beta$-delayed proton emission process which can be interpreted as the long sought coexisting 0$^+$ state., Comment: 8 pages with references, 3 figures. Supplemental material 4 pages (1 table, 3 figures)

The system of light electroweakinos and heavy squarks gives rise to one of the most challenging signatures to detect at the LHC. It consists of missing transverse energy recoiled against a few hadronic jets originating either from QCD radiation or squark decays. The analysis generally suffers from the large irreducible Z + jets $(Z \to \nu \bar \nu)$ background. In this study, we explore Machine Learning (ML) methods for efficient signal/background discrimination. Our best attempt uses both reconstructed (jets, missing transverse energy, etc.) and low-level (particle-flow) objects. We find that the discrimination performance improves as the pT threshold for soft particles is lowered from 10 GeV to 1 GeV, at the expense of larger systematic uncertainty. In many cases, the ML method provides a factor two enhancement in $S/\sqrt{(S + B)}$ from a simple kinematical selection. The sensitivity on the squark-elecroweakino mass plane is derived with this method, assuming the Run-3 and HL-LHC luminosities. Moreover, we investigate the relations between input features and the network's classification performance to reveal the physical information used in the background/signal discrimination process., Comment: 38 pages, 32 figures

We present the first demonstration of simultaneous phase space compression in two spatial dimensions of a positive muon beam, the first stage of the novel high-brightness muon beam under development by the muCool collaboration at the Paul Scherrer Institute. The keV-energy, sub-mm size beam would enable a factor 10$^5$ improvement in brightness for precision muSR, and atomic and particle physics measurements with positive muons. This compression is achieved within a cryogenic helium gas target with a strong density gradient, placed in a homogeneous magnetic field, under the influence of a complex electric field. In the next phase, the muon beam will be extracted into vacuum., Comment: Submission to SciPost

Recently, one of the present authors noticed a stimulated emission process of bosonic dark matter via the two-body decay of a mother particle in a thermal plasma similar to the operation principle of a laser in 2301.08735. In this paper, we show that in a $2 \to 2$ process, including a bosonic final particle (e.g., an axion or dark photon), the stimulated emission occurs as well due to a small angle scattering of the thermal mother particles and thus the phenomenon is more universal. Two important conclusions follow: (1) Care must be taken when studying the freeze-in production of a bosonic dark matter, as the abundance and momentum distribution of dark matter can differ significantly due to this effect. (2) eV-mass-range bosonic dark matter is special and theoretically well-motivated because models for freeze-in or other thermal production of dark matter include the parameter region of cold eV dark matter. We also study the dark matter mass effect for the stimulated emission., Comment: 15pages, 8 figures, Comments Wellcome!

Imaging atmospheric Cherenkov telescopes (IACTs) are used to observe very high-energy photons from the ground. Gamma rays are indirectly detected through the Cherenkov light emitted by the air showers they induce. The new generation of experiments, in particular the Cherenkov Telescope Array Observatory (CTAO), sets ambitious goals for discoveries of new gamma-ray sources and precise measurements of the already discovered ones. To achieve these goals, both hardware and data analysis must employ cutting-edge techniques. This also applies to the LST-1, the first IACT built for the CTAO, which is currently taking data on the Canary island of La Palma. This paper introduces a new event reconstruction technique for IACT data, aiming to improve the image reconstruction quality and the discrimination between the signal and the background from misidentified hadrons and electrons. The technique models the development of the extensive air shower signal, recorded as a waveform per pixel, seen by CTAO telescopes' cameras. Model parameters are subsequently passed to random forest regressors and classifiers to extract information on the primary particle. The new reconstruction was applied to simulated data and to data from observations of the Crab Nebula performed by the LST-1. The event reconstruction method presented here shows promising performance improvements. The angular and energy resolution, and the sensitivity, are improved by 10 to 20% over most of the energy range. At low energy, improvements reach up to 22%, 47%, and 50%, respectively. A future extension of the method to stereoscopic analysis for telescope arrays will be the next important step., Comment: Accepted in A&A

Retrieving gene functional networks from knowledge databases presents a challenge due to the mismatch between disease networks and subtype-specific variations. Current solutions, including statistical and deep learning methods, often fail to effectively integrate gene interaction knowledge from databases or explicitly learn subtype-specific interactions. To address this mismatch, we propose GeSubNet, which learns a unified representation capable of predicting gene interactions while distinguishing between different disease subtypes. Graphs generated by such representations can be considered subtype-specific networks. GeSubNet is a multi-step representation learning framework with three modules: First, a deep generative model learns distinct disease subtypes from patient gene expression profiles. Second, a graph neural network captures representations of prior gene networks from knowledge databases, ensuring accurate physical gene interactions. Finally, we integrate these two representations using an inference loss that leverages graph generation capabilities, conditioned on the patient separation loss, to refine subtype-specific information in the learned representation. GeSubNet consistently outperforms traditional methods, with average improvements of 30.6%, 21.0%, 20.1%, and 56.6% across four graph evaluation metrics, averaged over four cancer datasets. Particularly, we conduct a biological simulation experiment to assess how the behavior of selected genes from over 11,000 candidates affects subtypes or patient distributions. The results show that the generated network has the potential to identify subtype-specific genes with an 83% likelihood of impacting patient distribution shifts. The GeSubNet resource is available: https://anonymous.4open.science/r/GeSubNet/, Comment: Under review as a conference paper at ICLR 2025

While end-to-end multi-channel electroencephalography (EEG) learning approaches have shown significant promise, their applicability is often constrained in neurological diagnostics, such as intracranial EEG resources. When provided with a single-channel EEG, how can we learn representations that are robust to multi-channels and scalable across varied tasks, such as seizure prediction? In this paper, we present SplitSEE, a structurally splittable framework designed for effective temporal-frequency representation learning in single-channel EEG. The key concept of SplitSEE is a self-supervised framework incorporating a deep clustering task. Given an EEG, we argue that the time and frequency domains are two distinct perspectives, and hence, learned representations should share the same cluster assignment. To this end, we first propose two domain-specific modules that independently learn domain-specific representation and address the temporal-frequency tradeoff issue in conventional spectrogram-based methods. Then, we introduce a novel clustering loss to measure the information similarity. This encourages representations from both domains to coherently describe the same input by assigning them a consistent cluster. SplitSEE leverages a pre-training-to-fine-tuning framework within a splittable architecture and has following properties: (a) Effectiveness: it learns representations solely from single-channel EEG but has even outperformed multi-channel baselines. (b) Robustness: it shows the capacity to adapt across different channels with low performance variance. Superior performance is also achieved with our collected clinical dataset. (c) Scalability: With just one fine-tuning epoch, SplitSEE achieves high and stable performance using partial model layers., Comment: This paper has been accepted by ICDM2024

Recent normalization-based methods have shown great success in tackling the distribution shift issue, facilitating non-stationary time series forecasting. Since these methods operate in the time domain, they may fail to fully capture the dynamic patterns that are more apparent in the frequency domain, leading to suboptimal results. This paper first theoretically analyzes how normalization methods affect frequency components. We prove that the current normalization methods that operate in the time domain uniformly scale non-zero frequencies, and thus, they struggle to determine components that contribute to more robust forecasting. Therefore, we propose FredNormer, which observes datasets from a frequency perspective and adaptively up-weights the key frequency components. To this end, FredNormer consists of two components: a statistical metric that normalizes the input samples based on their frequency stability and a learnable weighting layer that adjusts stability and introduces sample-specific variations. Notably, FredNormer is a plug-and-play module, which does not compromise the efficiency compared to existing normalization methods. Extensive experiments show that FredNormer improves the averaged MSE of backbone forecasting models by 33.3% and 55.3% on the ETTm2 dataset. Compared to the baseline normalization methods, FredNormer achieves 18 top-1 results and 6 top-2 results out of 28 settings.

Recently, federated learning has attracted much attention as a privacy-preserving integrated analysis that enables integrated analysis of data held by multiple institutions without sharing raw data. On the other hand, federated learning requires iterative communication across institutions and has a big challenge for implementation in situations where continuous communication with the outside world is extremely difficult. In this study, we propose a federated data collaboration learning (FedDCL), which solves such communication issues by combining federated learning with recently proposed non-model share-type federated learning named as data collaboration analysis. In the proposed FedDCL framework, each user institution independently constructs dimensionality-reduced intermediate representations and shares them with neighboring institutions on intra-group DC servers. On each intra-group DC server, intermediate representations are transformed to incorporable forms called collaboration representations. Federated learning is then conducted between intra-group DC servers. The proposed FedDCL framework does not require iterative communication by user institutions and can be implemented in situations where continuous communication with the outside world is extremely difficult. The experimental results show that the performance of the proposed FedDCL is comparable to that of existing federated learning., Comment: 18 pages, 6 figures, 3 tables

In recent years, considerable research has been conducted on vision-language models that handle both image and text data; these models are being applied to diverse downstream tasks, such as "image-related chat," "image recognition by instruction," and "answering visual questions." Vision-language models (VLMs), such as Contrastive Language-Image Pre-training (CLIP), are also high-performance image classifiers that are being developed into domain adaptation methods that can utilize language information to extend into unseen domains. However, because these VLMs embed images as a single point in a unified embedding space, there is room for improvement in the classification accuracy. Therefore, in this study, we proposed the Latent Augmentation using Regional Embedding (LARE), which embeds the image as a region in the unified embedding space learned by the VLM. By sampling the augmented image embeddings from within this latent region, LARE enables data augmentation to various unseen domains, not just to specific unseen domains. LARE achieves robust image classification for domains in and out using augmented image embeddings to fine-tune VLMs. We demonstrate that LARE outperforms previous fine-tuning models in terms of image classification accuracy on three benchmarks. We also demonstrate that LARE is a more robust and general model that is valid under multiple conditions, such as unseen domains, small amounts of data, and imbalanced data., Comment: 10 pages, 4 figures

Backdoor attacks pose a significant threat to deep neural networks, particularly as recent advancements have led to increasingly subtle implantation, making the defense more challenging. Existing defense mechanisms typically rely on an additional clean dataset as a standard reference and involve retraining an auxiliary model or fine-tuning the entire victim model. However, these approaches are often computationally expensive and not always feasible in practical applications. In this paper, we propose a novel and lightweight defense mechanism, termed PAD-FT, that does not require an additional clean dataset and fine-tunes only a very small part of the model to disinfect the victim model. To achieve this, our approach first introduces a simple data purification process to identify and select the most-likely clean data from the poisoned training dataset. The self-purified clean dataset is then used for activation clipping and fine-tuning only the last classification layer of the victim model. By integrating data purification, activation clipping, and classifier fine-tuning, our mechanism PAD-FT demonstrates superior effectiveness across multiple backdoor attack methods and datasets, as confirmed through extensive experimental evaluation.

Deep neural networks (DNNs) have achieved significant success in real-world applications. However, safeguarding their intellectual property (IP) remains extremely challenging. Existing DNN watermarking for IP protection often require modifying DNN models, which reduces model performance and limits their practicality. This paper introduces FreeMark, a novel DNN watermarking framework that leverages cryptographic principles without altering the original host DNN model, thereby avoiding any reduction in model performance. Unlike traditional DNN watermarking methods, FreeMark innovatively generates secret keys from a pre-generated watermark vector and the host model using gradient descent. These secret keys, used to extract watermark from the model's activation values, are securely stored with a trusted third party, enabling reliable watermark extraction from suspect models. Extensive experiments demonstrate that FreeMark effectively resists various watermark removal attacks while maintaining high watermark capacity.

Multivariate time series data suffer from the problem of missing values, which hinders the application of many analytical methods. To achieve the accurate imputation of these missing values, exploiting inter-correlation by employing the relationships between sequences (i.e., a network) is as important as the use of temporal dependency, since a sequence normally correlates with other sequences. Moreover, exploiting an adequate network depending on time is also necessary since the network varies over time. However, in real-world scenarios, we normally know neither the network structure nor when the network changes beforehand. Here, we propose a missing value imputation method for multivariate time series, namely MissNet, that is designed to exploit temporal dependency with a state-space model and inter-correlation by switching sparse networks. The network encodes conditional independence between features, which helps us understand the important relationships for imputation visually. Our algorithm, which scales linearly with reference to the length of the data, alternatively infers networks and fills in missing values using the networks while discovering the switching of the networks. Extensive experiments demonstrate that MissNet outperforms the state-of-the-art algorithms for multivariate time series imputation and provides interpretable results., Comment: Accepted by KDD 2024

Machine learning has shown great potential in the field of cancer multi-omics studies, offering incredible opportunities for advancing precision medicine. However, the challenges associated with dataset curation and task formulation pose significant hurdles, especially for researchers lacking a biomedical background. Here, we introduce the CMOB, the first large-scale cancer multi-omics benchmark integrates the TCGA platform, making data resources accessible and usable for machine learning researchers without significant preparation and expertise.To date, CMOB includes a collection of 20 cancer multi-omics datasets covering 32 cancers, accompanied by a systematic data processing pipeline. CMOB provides well-processed dataset versions to support 20 meaningful tasks in four studies, with a collection of benchmarks. We also integrate CMOB with two complementary resources and various biological tools to explore broader research avenues.All resources are open-accessible with user-friendly and compatible integration scripts that enable non-experts to easily incorporate this complementary information for various tasks. We conduct extensive experiments on selected datasets to offer recommendations on suitable machine learning baselines for specific applications. Through CMOB, we aim to facilitate algorithmic advances and hasten the development, validation, and clinical translation of machine-learning models for personalized cancer treatments. CMOB is available on GitHub (\url{https://github.com/chenzRG/Cancer-Multi-Omics-Benchmark}).