Search

Asian Americans as a group are overrepresented among STEM college graduates and have the highest average college enrollment rate of any racial or ethnic category. Thus, Asian Americans are typically excluded from educational interventions directed at improving STEM education for Students of Color because they are not considered to be underrepresented minorities. However, statistics obscure the individual needs of the more than 20 ethnic subgroups that fall under the umbrella term Asian Americans. Using a participatory action research approach, this paper documents the institutional and sociocultural factors that push out HMoob (or Hmong) American college students from STEM programs at one large, predominantly White university; and the coordinate processes of gatekeeping and transactional advising that either redirect those students toward non-STEM programs or force them out of the university completely.

Evaluating the quality of synthesized images remains a significant challenge in the development of text-to-image (T2I) generation. Most existing studies in this area primarily focus on evaluating text-image alignment, image quality, and object composition capabilities, with comparatively fewer studies addressing the evaluation of the factuality of T2I models, particularly when the concepts involved are knowledge-intensive. To mitigate this gap, we present T2I-FactualBench in this work - the largest benchmark to date in terms of the number of concepts and prompts specifically designed to evaluate the factuality of knowledge-intensive concept generation. T2I-FactualBench consists of a three-tiered knowledge-intensive text-to-image generation framework, ranging from the basic memorization of individual knowledge concepts to the more complex composition of multiple knowledge concepts. We further introduce a multi-round visual question answering (VQA) based evaluation framework to assess the factuality of three-tiered knowledge-intensive text-to-image generation tasks. Experiments on T2I-FactualBench indicate that current state-of-the-art (SOTA) T2I models still leave significant room for improvement.

DarkSHINE is a newly proposed fixed-target experiment initiative to search for the invisible decay of Dark Photon via missing energy/momentum signatures, based on the high repetition rate electron beam to be deployed/delivered by the Shanghai High repetition rate XFEL and Extreme light facility (SHINE). This report elaborates the baseline design of DarkSHINE experiment by introducing the physics goals, experimental setups, details of each sub-detector system technical designs, signal and backgground modelings, expected search sensitivities and future prospects, which mark an important step towards the further prototyping and technical demonstrations.

We study the three-dimensional Carrollian field theory on the Rindler horizon which is dual to a bulk massless scalar field theory in the four-dimensional Rindler wedge. The Carrollian field theory could be mapped to a two-dimensional Euclidean field theory in the transverse plane by a Fourier transform. After defining the incoming and outgoing states at the future and past Rindler horizon, respectively, we construct the boundary-to-boundary and bulk-to-boundary propagators that are consistent with the bulk Green's function in the literature. We investigate the tree-level Carrollian amplitudes up to four points. The tree-level four-point Carrollian amplitude in $\Phi^4$ theory has the same structure as the one-loop triangle Feynman integral in the Lee-Pomeransky representation with complex powers in the propagators and spacetime dimension. Moreover, the four-point Carrollian amplitude with a zero energy state inserted at infinity in $\Phi^4$ theory is proportional to the three-point Carrollian amplitude in $\Phi^3$ theory., Comment: 78 pages, 14 figures, 8 tables

Employing aberration-corrected scanning transmission electron microscopy (STEM), we meticulously investigated the intrinsic chemical heterogeneity of Fe1+yTe, Fe1+yTe0.8Se0.2, and Fe1+yTe0.5Se0.5. Comprehensive analysis reveals the presence of interstitial iron atoms (Feint) across all samples, pre-dominantly occupying the 2c site of the P4/nmm space group. Moreover, a superstructure phase characterized by a wave vector q = 2/5a + 1/2c, originating from the ordering of Feint, is distinctly observable in the parent compound Fe1+yTe. In this scenario, the Feint atoms interact with adjacent Fe atoms, forming iron polycomplex and leading to an evident distortion of the FeTe4 tetrahedral. Experimental results further demonstrate effective suppression of Feint concentration and ordering through appropriate Se substitution; notably, Fe1+yTe0.5Se0.5 manifests the lowest concentration of Feint atoms. Our findings additionally indicate that Se substitution is random, and nanoscale phase separation induced by Te/Se chemical heterogeneity is commonly observed within Fe1+yTe1-xSex crystals., Comment: 14 pages, 5 figures

The emergence of Software-Defined Vehicles (SDVs) signifies a shift from a distributed network of electronic control units (ECUs) to a centralized computing architecture within the vehicle's electrical and electronic systems. This transition addresses the growing complexity and demand for enhanced functionality in traditional E/E architectures, with containerization and virtualization streamlining software development and updates within the SDV framework. While widely used in cloud computing, their performance and suitability for intelligent vehicles have yet to be thoroughly evaluated. In this work, we conduct a comprehensive performance evaluation of containerization and virtualization on embedded and high-performance AMD64 and ARM64 systems, focusing on CPU, memory, network, and disk metrics. In addition, we assess their impact on real-world automotive applications using the Autoware framework and further integrate a microservice-based architecture to evaluate its start-up time and resource consumption. Our extensive experiments reveal a slight 0-5% performance decline in CPU, memory, and network usage for both containerization and virtualization compared to bare-metal setups, with more significant reductions in disk operations-5-15% for containerized environments and up to 35% for virtualized setups. Despite these declines, experiments with actual vehicle applications demonstrate minimal impact on the Autoware framework, and in some cases, a microservice architecture integration improves start-up time by up to 18%., Comment: 15 pages, 15 figures

Anisotropic stochastic gravitational wave background (SGWB) serves as a potential probe of the large-scale structure (LSS) of the universe. In this work, we explore the anisotropic SGWB from local ($z < \sim 0.085$) merging stellar mass compact binaries, specifically focusing on merging stellar binary black holes, merging neutron-star-black-hole binaries, and merging binary neutron stars. The analysis employs seven all-sky mock lightcone gravitational wave event catalogues, which are derived from the Millennium simulation combined with a semi-analytic model of galaxy formation and a binary population synthesis model. We calculate the angular power spectra $\mathrm{C}_\ell$ at multipole moments $\ell$, expressed as $\text{log}_{10} [\ell(\ell+1)\mathrm{C}_\ell/(2\pi)]$, based on the skymaps of the overdensity $\delta_\mathrm{GW}$ in the anisotropic SGWB. The spectra for all three source types exhibit an approximately linear increase with $\text{log}_{10} \ell$ at higher $\ell$ (e.g., $\ell > \sim 30 - 300$) in seven catalogues, with a characteristic slope of $\sim 2$. The spectra of seven catalogues exhibit considerable variations, arising from fluctuations in spatial distribution, primarily in the radial distribution, of nearby sources (e.g., $< 50$ Mpc/h). After subtracting these nearby sources, the variations become much smaller and the spectra for the three source types become closely aligned (within discrepancies of a factor of $\sim 2$ across $\ell = 1 - 1000$ for all catalogues)., Comment: 11 pages 5 figures

Cognitive impairment detection through spontaneous speech offers potential for early diagnosis of Alzheimer's disease (AD) and mild cognitive impairment (MCI). The PROCESS Grand Challenge, part of ICASSP 2025, focuses on advancing this field with innovative solutions for classification and regression tasks. In this work, we integrate interpretable features with temporal features extracted from pre-trained models through a multimodal fusion strategy. For the classification task, our model achieved an F1-score of 0.649 in predicting cognitive states (healthy, MCI, dementia). For the regression task, which involves MMSE score prediction, we obtained a root-mean-square error (RMSE) of 2.628. These results led to our team securing the top overall ranking in the competition., Comment: ICASSP 2025

Deep hash-based retrieval techniques are widely used in facial retrieval systems to improve the efficiency of facial matching. However, it also brings the risk of privacy leakage. Deep hash models are easily influenced by adversarial examples, which can be leveraged to prevent the malicious retrieval of private images. The existing adversarial example methods against deep hash models focus on universality and transferability, lacking the research on its robustness in online social networks (OSNs), which leads to their failure in anti-retrieval after post-processing. Therefore, we provide the first in-depth discussion on robustness adversarial perturbation in universal transferable anti-facial retrieval and propose Three-in-One Adversarial Perturbation (TOAP). Specifically, we firstly analyze the performance of deep hash models after post-processing and construct a local and global Compression Generator (CG) to simulate complex post-processing scenarios. Then, we explore the variation patterns of the model's objective under image post-processing and propose robust optimization objectives, cluster centers and data space centers, optimizing them using meta-learning. Finally, we iteratively optimize perturbation by alternately generating adversarial examples and fine-tuning the CG, balancing the performance of perturbation while enhancing CG's ability to mitigate them. Numerous experiments demonstrate that, in addition to its advantages in universality and transferability, TOAP significantly outperforms current state-of-the-art methods in multiple robustness metrics. It further improves universality and transferability by 5% to 28%, and achieves up to about 33% significant improvement in several simulated post-processing scenarios as well as mainstream OSNs, demonstrating that TOAP can effectively protect private images from malicious retrieval in real-world scenarios.

We explore a novel video creation experience, namely Video Creation by Demonstration. Given a demonstration video and a context image from a different scene, we generate a physically plausible video that continues naturally from the context image and carries out the action concepts from the demonstration. To enable this capability, we present $\delta$-Diffusion, a self-supervised training approach that learns from unlabeled videos by conditional future frame prediction. Unlike most existing video generation controls that are based on explicit signals, we adopts the form of implicit latent control for maximal flexibility and expressiveness required by general videos. By leveraging a video foundation model with an appearance bottleneck design on top, we extract action latents from demonstration videos for conditioning the generation process with minimal appearance leakage. Empirically, $\delta$-Diffusion outperforms related baselines in terms of both human preference and large-scale machine evaluations, and demonstrates potentials towards interactive world simulation. Sampled video generation results are available at https://delta-diffusion.github.io/., Comment: Project page at https://delta-diffusion.github.io/

As information becomes more accessible, user-generated videos are increasing in length, placing a burden on viewers to sift through vast content for valuable insights. This trend underscores the need for an algorithm to extract key video information efficiently. Despite significant advancements in highlight detection, moment retrieval, and video summarization, current approaches primarily focus on selecting specific time intervals, often overlooking the relevance between segments and the potential for segment arranging. In this paper, we introduce a novel task called Video Trimming (VT), which focuses on detecting wasted footage, selecting valuable segments, and composing them into a final video with a coherent story. To address this task, we propose Agent-based Video Trimming (AVT), structured into three phases: Video Structuring, Clip Filtering, and Story Composition. Specifically, we employ a Video Captioning Agent to convert video slices into structured textual descriptions, a Filtering Module to dynamically discard low-quality footage based on the structured information of each clip, and a Video Arrangement Agent to select and compile valid clips into a coherent final narrative. For evaluation, we develop a Video Evaluation Agent to assess trimmed videos, conducting assessments in parallel with human evaluations. Additionally, we curate a new benchmark dataset for video trimming using raw user videos from the internet. As a result, AVT received more favorable evaluations in user studies and demonstrated superior mAP and precision on the YouTube Highlights, TVSum, and our own dataset for the highlight detection task. The code and models are available at https://ylingfeng.github.io/AVT.

We revisit the launch of the galactic outflow in M82 through high-resolution hydrodynamic simulations. Employing a sink-particle module, we self-consistently resolve the star formation and feedback processes, avoiding the reliance on various assumed models as in previous works. We probe the effects of different stellar feedback mechanisms, the gas return from star-forming clouds and initial gas disc mass on the starburst and launch of outflow. Our simulations can generate a starburst that lasts 20-30 Myr, peaking at 20-50 $\rm{M_{\odot} yr^{-1}}$. However, the total stellar mass formed in the starburst in our simulations often exceeds M82's estimated value. The outflow's launch occurs in two stages. Initially, continuous SNe explosions form numerous small bubbles, merging into a super bubble foam composed of low-density warm/hot gas and high-density cool filaments. After approximately 10 Myr of SN energy injection, the super bubble breakout the disc, marking the second stage, which takes 10-15 Myr to develop a multiphase, kpc-scale outflow. Our simulations reveal that cool filaments within the turbulent ISM can survive from the feedback associated to the starburst, then were entrained into the outflow and stretched to hundreds pc in length due to interaction with the hot wind. While the mass loading factor of the well-developed outflow is comparable to M82, the cool gas mass outflow rate is often lower, and its velocity is slower than the estimated value in M82 by $50\%-60\%$. Warm and hot gas are $20\%-30\%$ slower. SN feedback acts as the primary driver of the outflow, while gas return significantly influences the starburst and outflow properties. The initial mass of gas disc has a moderate effect, while stellar wind and radiation feedback have minor effect. To address the shortcomings in our results, enhanced SN feedback effect, due to clustered SNe, is likely necessary., Comment: 26 pages, 15 figures, to be submitted. Comments welcome!

Class-Incremental Learning (CIL) requires models to continually acquire knowledge of new classes without forgetting old ones. Despite Pre-trained Models (PTMs) have shown excellent performance in CIL, catastrophic forgetting still occurs as the model learns new concepts. Existing work seeks to utilize lightweight components to adjust the PTM, while the forgetting phenomenon still comes from {\em parameter and retrieval} levels. Specifically, iterative updates of the model result in parameter drift, while mistakenly retrieving irrelevant modules leads to the mismatch during inference. To this end, we propose MOdel Surgery (MOS) to rescue the model from forgetting previous knowledge. By training task-specific adapters, we continually adjust the PTM to downstream tasks. To mitigate parameter-level forgetting, we present an adapter merging approach to learn task-specific adapters, which aims to bridge the gap between different components while reserve task-specific information. Besides, to address retrieval-level forgetting, we introduce a training-free self-refined adapter retrieval mechanism during inference, which leverages the model's inherent ability for better adapter retrieval. By jointly rectifying the model with those steps, MOS can robustly resist catastrophic forgetting in the learning process. Extensive experiments on seven benchmark datasets validate MOS's state-of-the-art performance. Code is available at: https://github.com/sun-hailong/AAAI25-MOS, Comment: Accepted to AAAI 2025. Code is available at: https://github.com/sun-hailong/AAAI25-MOS

A search for $D^0$ meson decays to the $\pi^+\pi^-e^+e^-$ and $K^+K^-e^+e^-$ final states is reported using a sample of proton-proton collisions collected by the LHCb experiment at a center-of-mass energy of 13 TeV, corresponding to an integrated luminosity of 6 fb$^{-1}$. The decay $D^0 \rightarrow \pi^+\pi^-e^+e^-$ is observed for the first time when requiring that the two electrons are consistent with coming from the decay of a $\phi$ or $\rho^0/\omega$ meson. The corresponding branching fractions are measured relative to the $D^0 \rightarrow K^-\pi^-[e^+e^-]_{\rho^0/\omega}$ decay, where the two electrons are consistent with coming from the decay of a $\rho^0$ or $\omega$ meson. No evidence is found for the $D^0 \rightarrow K^+K^-e^+e^-$ decay and world-best limits are set on its branching fraction. The results are compared to, and found to be consistent with, the branching fractions of the $D^0 \rightarrow \pi^+\pi^-\mu^+\mu^-$ and $D^0 \rightarrow K^+K^-\mu^+\mu^-$ decays recently measured by LHCb and confirm lepton universality at the current precision., Comment: All figures and tables, along with machine-readable versions and any supplementary material and additional information, are available at https://lbfence.cern.ch/alcm/public/analysis/full-details/1611/ (LHCb public pages)

Controllable human image animation aims to generate videos from reference images using driving videos. Due to the limited control signals provided by sparse guidance (e.g., skeleton pose), recent works have attempted to introduce additional dense conditions (e.g., depth map) to ensure motion alignment. However, such strict dense guidance impairs the quality of the generated video when the body shape of the reference character differs significantly from that of the driving video. In this paper, we present DisPose to mine more generalizable and effective control signals without additional dense input, which disentangles the sparse skeleton pose in human image animation into motion field guidance and keypoint correspondence. Specifically, we generate a dense motion field from a sparse motion field and the reference image, which provides region-level dense guidance while maintaining the generalization of the sparse pose control. We also extract diffusion features corresponding to pose keypoints from the reference image, and then these point features are transferred to the target pose to provide distinct identity information. To seamlessly integrate into existing models, we propose a plug-and-play hybrid ControlNet that improves the quality and consistency of generated videos while freezing the existing model parameters. Extensive qualitative and quantitative experiments demonstrate the superiority of DisPose compared to current methods. Code: \href{https://github.com/lihxxx/DisPose}{https://github.com/lihxxx/DisPose}.

Bayesian Optimisation (BO) is a state-of-the-art global optimisation technique for black-box problems where derivative information is unavailable, and sample efficiency is crucial. However, improving the general scalability of BO has proved challenging. Here, we explore Latent Space Bayesian Optimisation (LSBO), that applies dimensionality reduction to perform BO in a reduced-dimensional subspace. While early LSBO methods used (linear) random projections (Wang et al., 2013), we employ Variational Autoencoders (VAEs) to manage more complex data structures and general DR tasks. Building on Grosnit et. al. (2021), we analyse the VAE-based LSBO framework, focusing on VAE retraining and deep metric loss. We suggest a few key corrections in their implementation, originally designed for tasks such as molecule generation, and reformulate the algorithm for broader optimisation purposes. Our numerical results show that structured latent manifolds improve BO performance. Additionally, we examine the use of the Mat\'{e}rn-$\frac{5}{2}$ kernel for Gaussian Processes in this LSBO context. We also integrate Sequential Domain Reduction (SDR), a standard global optimization efficiency strategy, into BO. SDR is included in a GPU-based environment using \textit{BoTorch}, both in the original and VAE-generated latent spaces, marking the first application of SDR within LSBO., Comment: Accepted at NeurIPS 2024 Workshop OPT for ML: Optimization for Machine Learning (Submission Number:67)

A simple orbit classification constraint extension to stellar dynamical modeling using Schwarzschild's method is demonstrated. The classification scheme used is the existing `orbit circularity' scheme (lambda_z) where orbits are split into four groups - hot, warm, cold and counter rotating orbits. Other schemes which can be related to the orbit weights are expected to be viable as well. The results show that the classification constraint works well in modeling. However, given that orbits in external galaxies are not observable, it is not clear how the orbit classification for any particular galaxy may be determined. Perhaps range constraints for different types of galaxies determined from cosmological simulations may offer a way forward., Comment: Accepted for publication, 15 pages, 3 figures, 9 tables

Text embedding has become a foundational technology in natural language processing (NLP) during the deep learning era, driving advancements across a wide array of downstream tasks. While many natural language understanding challenges can now be modeled using generative paradigms and leverage the robust generative and comprehension capabilities of large language models (LLMs), numerous practical applications, such as semantic matching, clustering, and information retrieval, continue to rely on text embeddings for their efficiency and effectiveness. In this survey, we categorize the interplay between LLMs and text embeddings into three overarching themes: (1) LLM-augmented text embedding, enhancing traditional embedding methods with LLMs; (2) LLMs as text embedders, utilizing their innate capabilities for embedding generation; and (3) Text embedding understanding with LLMs, leveraging LLMs to analyze and interpret embeddings. By organizing these efforts based on interaction patterns rather than specific downstream applications, we offer a novel and systematic overview of contributions from various research and application domains in the era of LLMs. Furthermore, we highlight the unresolved challenges that persisted in the pre-LLM era with pre-trained language models (PLMs) and explore the emerging obstacles brought forth by LLMs. Building on this analysis, we outline prospective directions for the evolution of text embedding, addressing both theoretical and practical opportunities in the rapidly advancing landscape of NLP., Comment: Work in progress

The requirements of unitarity and causality lead to significant constraints on the Wilson coefficients of a EFT expansion, known as positivity bounds. Their standard derivation relies on the crucial assumption of polynomial boundedness on the growth of scattering amplitudes in the complex energy plane, which is a property satisfied by local QFTs, and by weakly coupled string theory in the Regge regime. The scope of this work is to clarify the role of locality by deriving generalized positivity bounds under the assumption of exponential boundedness, typical of non-local QFTs where the Froissart-Martin bound is usually not satisfied. Using appropriately modified dispersion relations, we derive new constraints and find regions in the EFT parameter space that do not admit a local UV completion. Furthermore, we show that there exist ETFs that satisfy IR causality and at the same time can admit a non-local UV completion, provided that the energy scale of non-locality is of the same order or smaller than the EFT cutoff. Finally, we provide explicit examples of non-perturbative amplitudes that simultaneously satisfy the properties of exponential boundedness, unitarity and causality. Our results have far-reaching implications for the question of the uniqueness of string theory as the only consistent ultraviolet completion beyond the framework of local QFT., Comment: 25 pages, 5 figures

Sequential recommendation systems aim to provide personalized recommendations for users based on their interaction history. To achieve this, they often incorporate auxiliary information, such as textual descriptions of items and auxiliary tasks, like predicting user preferences and intent. Despite numerous efforts to enhance these models, they still suffer from limited personalization. To address this issue, we propose a new paradigm, which we term preference discerning. In preference dscerning, we explicitly condition a generative sequential recommendation system on user preferences within its context. To this end, we generate user preferences using Large Language Models (LLMs) based on user reviews and item-specific data. To evaluate preference discerning capabilities of sequential recommendation systems, we introduce a novel benchmark that provides a holistic evaluation across various scenarios, including preference steering and sentiment following. We assess current state-of-the-art methods using our benchmark and show that they struggle to accurately discern user preferences. Therefore, we propose a new method named Mender ($\textbf{M}$ultimodal Prefer$\textbf{en}$ce $\textbf{d}$iscern$\textbf{er}$), which improves upon existing methods and achieves state-of-the-art performance on our benchmark. Our results show that Mender can be effectively guided by human preferences even though they have not been observed during training, paving the way toward more personalized sequential recommendation systems. We will open-source the code and benchmarks upon publication., Comment: 11 pages + references and appendix

In this paper, we give an explicit formula of the Igusa local zeta function of a Thom-Sebastiani type sum of two separated-variable Newton non-critical polynomials. Data for the description are available on their Newton polyhedra., Comment: 17 pages

In the $N=1$ superspace, AdS$_4$ supersymmetry is realized as the non-linear super coordinate transformations. The fermionic coordinates form a faithful non-linear representation of supersymmetry on their own. By introducing an auxiliary scalar coordinate, this representation is reformulated as a 5-dimensional linear representation, i.e., the superspinor representation. New linear representations are constructed by tensor products of multiple superspinors. Especially, the superspace bosonic coordinates are embedded in the supervector representation, which is the traceless symmetric part of the bi-superspinor representation. Based on these linear representations, the $N=1$ AdS$_4$ supergravity action can be reproduced in a manifestly supersymmetry-covariant way., Comment: 9 pages, no figures

Maximal clique enumeration (MCE) is crucial for tasks like community detection and biological network analysis. Existing algorithms typically adopt the branch-and-bound framework with the vertex-oriented Bron-Kerbosch (BK) branching strategy, which forms the sub-branches by expanding the partial clique with a vertex. In this paper, we present a novel approach called HBBMC, a hybrid framework combining vertex-oriented BK branching and edge-oriented BK branching, where the latter adopts a branch-and-bound framework which forms the sub-branches by expanding the partial clique with an edge. This hybrid strategy enables more effective pruning and helps achieve a worst-case time complexity better than the best known one under a condition that holds for the majority of real-world graphs. To further enhance efficiency, we introduce an early termination technique, which leverages the topological information of the graphs and constructs the maximal cliques directly without branching. Our early termination technique is applicable to all branch-and-bound frameworks. Extensive experiments demonstrate the superior performance of our techniques., Comment: Accepted by ICDE'25

Existing Large Vision-Language Models (LVLMs) excel at matching concepts across multi-modal inputs but struggle with compositional concepts and high-level relationships between entities. This paper introduces Progressive multi-granular Vision-Language alignments (PromViL), a novel framework to enhance LVLMs' ability in performing grounded compositional visual reasoning tasks. Our approach constructs a hierarchical structure of multi-modal alignments, ranging from simple to complex concepts. By progressively aligning textual descriptions with corresponding visual regions, our model learns to leverage contextual information from lower levels to inform higher-level reasoning. To facilitate this learning process, we introduce a data generation process that creates a novel dataset derived from Visual Genome, providing a wide range of nested compositional vision-language pairs. Experimental results demonstrate that our PromViL framework significantly outperforms baselines on various visual grounding and compositional question answering tasks.

Training a general-purpose time series foundation models with robust generalization capabilities across diverse applications from scratch is still an open challenge. Efforts are primarily focused on fusing cross-domain time series datasets to extract shared subsequences as tokens for training models on Transformer architecture. However, due to significant statistical heterogeneity across domains, this cross-domain fusing approach doesn't work effectively as the same as fusing texts and images. To tackle this challenge, this paper proposes a novel federated learning approach to address the heterogeneity in time series foundation models training, namely FFTS. Specifically, each data-holding organization is treated as an independent client in a collaborative learning framework with federated settings, and then many client-specific local models will be trained to preserve the unique characteristics per dataset. Moreover, a new regularization mechanism will be applied to both client-side and server-side, thus to align the shared knowledge across heterogeneous datasets from different domains. Extensive experiments on benchmark datasets demonstrate the effectiveness of the proposed federated learning approach. The newly learned time series foundation models achieve superior generalization capabilities on cross-domain time series analysis tasks, including forecasting, imputation, and anomaly detection., Comment: Accepted by Main Track in AAAI'25

Frontier models are currently developed and distributed primarily through two channels: centralized proprietary APIs or open-sourcing of pre-trained weights. We identify a third paradigm - Protocol Learning - where models are trained across decentralized networks of incentivized participants. This approach has the potential to aggregate orders of magnitude more computational resources than any single centralized entity, enabling unprecedented model scales and capabilities. However, it also introduces novel challenges: heterogeneous and unreliable nodes, malicious participants, the need for unextractable models to preserve incentives, and complex governance dynamics. To date, no systematic analysis has been conducted to assess the feasibility of Protocol Learning or the associated risks, particularly the 'No-Off Problem' arising from the inability to unilaterally halt a collectively trained model. We survey recent technical advances that suggest decentralized training may be feasible - covering emerging communication-efficient strategies and fault-tolerant methods - while highlighting critical open problems that remain. Contrary to the notion that decentralization inherently amplifies frontier risks, we argue that Protocol Learning's transparency, distributed governance, and democratized access ultimately reduce these risks compared to today's centralized regimes.

The rapid growth of memecoins within the Web3 ecosystem, driven by platforms like Pump.fun, has made it easier for anyone to create tokens. However, this democratization has also led to an explosion of low-quality or bot-generated projects, often motivated by short-term financial gain. This overwhelming influx of speculative tokens creates a challenge in distinguishing viable memecoins from those that are unlikely to succeed. To address this issue, we introduce CoinVibe, a comprehensive multimodal dataset designed to evaluate the viability of memecoins. CoinVibe integrates textual descriptions, visual content (logos), and community data (user comments, timestamps, and number of likes) to provide a holistic view of a memecoin's potential. In addition, we present CoinCLIP, a novel framework that leverages the Contrastive Language-Image Pre-Training (CLIP) model, augmented with lightweight modules and community data integration, to improve classification accuracy. By combining visual and textual representations with community insights, CoinCLIP provides a robust, data-driven approach to filter out low-quality or bot-driven projects. This research aims to help creators and investors identify high-potential memecoins, while also offering valuable insights into the factors that contribute to their long-term success. The code and dataset are publicly available at https://github.com/hwlongCUHK/CoinCLIP.git., Comment: 4 pages, 1 figure, conference

Most existing mobile robotic datasets primarily capture static scenes, limiting their utility for evaluating robotic performance in dynamic environments. To address this, we present a mobile robot oriented large-scale indoor dataset, denoted as THUD++ (TsingHua University Dynamic) robotic dataset, for dynamic scene understanding. Our current dataset includes 13 large-scale dynamic scenarios, combining both real-world and synthetic data collected with a real robot platform and a physical simulation platform, respectively. The RGB-D dataset comprises over 90K image frames, 20M 2D/3D bounding boxes of static and dynamic objects, camera poses, and IMU. The trajectory dataset covers over 6,000 pedestrian trajectories in indoor scenes. Additionally, the dataset is augmented with a Unity3D-based simulation platform, allowing researchers to create custom scenes and test algorithms in a controlled environment. We evaluate state-of-the-art methods on THUD++ across mainstream indoor scene understanding tasks, e.g., 3D object detection, semantic segmentation, relocalization, pedestrian trajectory prediction, and navigation. Our experiments highlight the challenges mobile robots encounter in indoor environments, especially when navigating in complex, crowded, and dynamic scenes. By sharing this dataset, we aim to accelerate the development and testing of mobile robot algorithms, contributing to real-world robotic applications., Comment: arXiv admin note: substantial text overlap with arXiv:2406.19791

The security of AI-generated content (AIGC) detection based on GANs and diffusion models is closely related to the credibility of multimedia content. Malicious adversarial attacks can evade these developing AIGC detection. However, most existing adversarial attacks focus only on GAN-generated facial images detection, struggle to be effective on multi-class natural images and diffusion-based detectors, and exhibit poor invisibility. To fill this gap, we first conduct an in-depth analysis of the vulnerability of AIGC detectors and discover the feature that detectors vary in vulnerability to different post-processing. Then, considering the uncertainty of detectors in real-world scenarios, and based on the discovery, we propose a Realistic-like Robust Black-box Adversarial attack (R$^2$BA) with post-processing fusion optimization. Unlike typical perturbations, R$^2$BA uses real-world post-processing, i.e., Gaussian blur, JPEG compression, Gaussian noise and light spot to generate adversarial examples. Specifically, we use a stochastic particle swarm algorithm with inertia decay to optimize post-processing fusion intensity and explore the detector's decision boundary. Guided by the detector's fake probability, R$^2$BA enhances/weakens the detector-vulnerable/detector-robust post-processing intensity to strike a balance between adversariality and invisibility. Extensive experiments on popular/commercial AIGC detectors and datasets demonstrate that R$^2$BA exhibits impressive anti-detection performance, excellent invisibility, and strong robustness in GAN-based and diffusion-based cases. Compared to state-of-the-art white-box and black-box attacks, R$^2$BA shows significant improvements of 15% and 21% in anti-detection performance under the original and robust scenario respectively, offering valuable insights for the security of AIGC detection in real-world applications.

In recent years, non-equilibrium flows have gained significant attention in aerospace engineering and micro-electro-mechanical systems. The unified gas-kinetic scheme (UGKS) follows the methodology of direct modeling to couple particle collisions and free transport during gas evolution. However, like other discrete-velocity-based methods, the UGKS faces challenges related to high memory requirements and computational costs, such as the possible consumption of $1.32$ TB of memory when using $512$ cores for the simulations of the hypersonic flow around an X38-like space vehicle. This paper introduces a new UGKS programming paradigm for unstructured grids, focusing on reducing memory usage and improving parallel efficiency. By optimizing the computational sequence, the current method enables each cell in physical space to store only the distribution function for the discretized velocity space, eliminating the need to retain the entire velocity space for slopes and residuals. Additionally, the parallel communication is enhanced through the use of non-blocking MPI. Numerical experiments demonstrate that the new strategy in the programming effectively simulates non-equilibrium problems while achieving high computational efficiency and low memory consumption. For the hypersonic flow around an X38-like space vehicle, the simulation, which utilizes $1,058,685$ physical mesh cells and $4,548$ discrete velocity space mesh cells, requires only $168.12$ GB of memory when executed on $512$ CPU cores. This indicates that memory consumption in the UGKS is much reduced. This new programming paradigm can serve as a reference for discrete velocity methods for solving kinetic equations.

This study delves into the existence of dark matter halo around supermassive black holes in galactic cores using a novel gravitational model. By analyzing gravitational waves emitted during the ringdown phase of black holes under different field perturbations, we explore the potential for detecting dark matter halos. The model hypothesizes that the dark matter distribution around black hole is driven by a mechanism where dark energy endows gravitons with mass, thereby forming a new spacetime structure. Results reveal that as relevant parameters increase, the quasinormal modes (QNMs) exhibit a gradual reduction in real parts, with negative imaginary parts whose absolute values also decrease. Moreover, compared to gravitational wave signals from Schwarzschild black hole without dark matter halos, this system demonstrates significant differences in oscillation modes and frequencies. This achievement not only validates the self-consistency of the new gravitational model but also lays a theoretical foundation for subsequent gravitational wave detection within dark matter halo. Simultaneously, it provides new theoretical support for understanding the mechanism of dark energy in large-scale cosmic structures and broadens the research perspective on the relationships between black hole physics, dark matter, and dark energy.

Existing task-oriented AI agents often depend on explicit instructions or external rewards, limiting their ability to be driven by intrinsic motivations like humans. In this paper, we present a desire-driven autonomy framework to guide a Large Language Model-based (LLM-based) agent to simulate human-like daily activities. In contrast to previous agents, our Desire-driven Autonomous Agent (D2A) operates on the principle of intrinsic desire, allowing it to propose and select tasks that fulfill its motivational framework autonomously. Inspired by the Theory of Needs, the motivational framework incorporates an understanding of human-like desires, such as the need for social interaction, personal fulfillment, and self-care. Utilizing a desire-driven task generation mechanism, the agent evaluates its current state and takes a sequence of activities aligned with its intrinsic motivations. Through simulations, we demonstrate that our Desire-driven Autonomous Agent (D2A) generates coherent, contextually relevant daily activities while exhibiting variability and adaptability similar to human behavior. A comparative analysis with other LLM-based frameworks demonstrates that our approach significantly enhances the rationality of the simulated activities.

In recent years, a significant number of oxygen-bearing complex organic molecules (COMs) have been detected in the gas phase of cold dark clouds such as TMC-1. The formation of these COMs cannot be explained by diffusive mechanisms on grains and gas phase reactions. This study investigates the formation of oxygen-bearing COMs in cold dark clouds using multiphase gas-grain models that incorporate cosmic ray-induced non-diffusive radiation chemistry and non-thermal sputtering desorption mechanisms. Additionally, we present the effects of varying elemental C/O ratio and different sputtering rates. We utilized an accelerated Gillespie algorithm, based on the regular Gillespie algorithm. The results of our models for dimethyl ether (CH3OCH3), methyl formate (HCOOCH3), acetaldehyde (CH3CHO), ethanol (C2H5OH), and methanol (CH3OH) show reasonable agreement with observations toward TMC-1, within a factor of 3. Out of the 94 species compared with observations, 63 show agreement within 1 order of magnitude, accounting for 67.02%. Overall inclusion of non-thermal mechanisms in multi-phase models shows notable improvement of modeling on oxygen-bearing COMs in the interstellar medium.

Compared to other clinical screening techniques, speech-and-language-based automated Alzheimer's disease (AD) detection methods are characterized by their non-invasiveness, cost-effectiveness, and convenience. Previous studies have demonstrated the efficacy of fine-tuning pre-trained language models (PLMs) for AD detection. However, the objective of this traditional fine-tuning method, which involves inputting only transcripts, is inconsistent with the masked language modeling (MLM) task used during the pre-training phase of PLMs. In this paper, we investigate prompt-based fine-tuning of PLMs, converting the classification task into a MLM task by inserting prompt templates into the transcript inputs. We also explore the impact of incorporating pause information from forced alignment into manual transcripts. Additionally, we compare the performance of various automatic speech recognition (ASR) models and select the Whisper model to generate ASR-based transcripts for comparison with manual transcripts. Furthermore, majority voting and ensemble techniques are applied across different PLMs (BERT and RoBERTa) using different random seeds. Ultimately, we obtain maximum detection accuracy of 95.8% (with mean 87.9%, std 3.3%) using manual transcripts, achieving state-of-the-art performance for AD detection using only transcripts on the ADReSS test set., Comment: Accepted by ISCSLP 2024

Community search over bipartite graphs is a fundamental problem, and finding influential communities has attracted significant attention. However, all existing studies have used the minimum weight of vertices as the influence of communities. This leads to an inaccurate assessment of real influence in graphs where there are only a few vertices with low weights. In this paper, we propose a new cohesive subgraph model named ($\alpha$,$\beta$)-influential community that considers the average weight of vertices from two layers on bipartite graphs, thereby providing a more comprehensive reflection of community influence. Based on this community model, we present a recursive algorithm that traverses the entire bipartite graph to find top-$r$ ($\alpha$,$\beta$)-influential communities. To further expedite the search for influential communities, we propose a slim tree structure to reduce the search width and introduce several effective upper bounds to reduce the search depth. Since we have proven that this problem is NP-hard, using exact algorithms to find top-$r$ ($\alpha$,$\beta$)-communities accurately is very time-consuming. Therefore, we propose an approximate algorithm using a greedy approach to find top-$r$ ($\alpha$,$\beta$)-communities as quickly as possible. It only takes $O((n+m)+m\log_{}{n})$ time. Additionally, we introduce a new pruning algorithm to improve the efficiency of the search. Extensive experiments on 10 real-world graphs validate both the effectiveness and the efficiency of our algorithms.

This paper addresses the challenge of teleoperating continuum instruments for minimally invasive surgery (MIS). We develop and adopt a novel task-priority-based kinematic formulation to quantitatively investigate teleoperation commands for continuum instruments under remote center of motion (RCM) constraints. Using redundancy resolution methods, we investigate the kinematic performance during teleoperation, comparing linear and angular commands within a task-priority scheme. For experimental validation, an instrument module (IM) was designed and integrated with a 7-DoF manipulator. Assessments, simulations, and experimental validations demonstrated the effectiveness of the proposed framework. The experiments involved several tasks: trajectory tracking of the IM tip along multiple paths with varying priorities for linear and angular teleoperation commands, pushing a ball along predefined paths on a silicon board, following a pattern on a pegboard, and guiding the continuum tip through rings on a ring board using a standard surgical kit., Comment: 26 pages, submitted to ASME Journal of Mechanisms and Robotics, under review

Gravitational waves (GWs) from core-collapse supernovae (CCSNe) have been proposed as a means to probe the internal physical properties of supernovae. However, due to their complex time-frequency structure, effectively searching for and extracting GW signals from CCSNe remains an unsolved challenge. In this paper, we apply the improved multisynchrosqueezing transform (IMSST) method to reconstruct simulated GW data based on the advanced LIGO (aLIGO) and Einstein Telescope (ET) detectors. These data are generated by the magnetorotational and neutrino-driven mechanisms, and we use the match score as the criterion for evaluating the quality of the reconstruction. To assess whether the reconstructed waveforms correspond to true GW signals, we calculate the false alarm probability of reconstruction (FAPR). For GW sources located at 10 kpc and datasets where the waveform amplitudes are normalized to $5 \times 10^{-21}$ observed by aLIGO, FAPR are $2.1 \times 10^{-2}$ and $6.2 \times 10^{-3}$, respectively. For GW sources at 100 kpc and with waveform amplitudes normalized to $5 \times 10^{-21}$ observed by ET, FAPR are $1.3 \times 10^{-1}$ and $1.5 \times 10^{-2}$, respectively. When the gravitational wave strain reaches $7 \times 10^{-21}$ and the match score threshold is set to 0.75, the IMSST method achieves maximum reconstruction distances of approximately 37 kpc and 317 kpc for aLIGO and ET, respectively. Finally, we compared the performance of IMSST and STFT in waveform reconstruction based on the ET. The results show that the maximum reconstructable distance using STFT is 186 kpc.

This paper deals with a second-order primal-dual dynamical system with Hessian-driven damping and Tikhonov regularization terms in connection with a convex-concave bilinear saddle point problem. We first obtain a fast convergence rate of the primal-dual gap along the trajectory generated by the dynamical system, and provide some integral estimates. Then, based on the setting of the parameters involved, we demonstrate that both the convergence rate of the primal-dual gap and the strong convergence of the trajectory can be achieved simultaneously. Furthermore, we evaluate the performance of the proposed system using two numerical examples.

The study of extra charged gauge boson beyond the Standard Model has always been of great interest. Future muon colliders will have a significant advantage in discovering exotic particles. In this paper, by studying the $\mu^+ \mu^- \to W^{\prime +} W^{\prime -} \to e^+ e^- n_e \bar{n}_e$ process, we explore the properties of $W^\prime$ in the alternative left-right model. The cross section and angular distribution of the final electron are investigated in the scenario of different $W^\prime$ mass and right-handed coupling constant. The forward-backward asymmetry is also an important observable to reflect the properties of $W^\prime$. We provide a method to effectively suppress the background processes. With specific kinematic cuts, the significance can reach $5.17\sigma$ for 4.8 TeV $W^\prime$ at the collision energy of 10 TeV., Comment: 18 pages,6 figures

We report analysis of $\textit{Insight}$-HXMT observations of the newly discovered accreting millisecond pulsar SRGA J144459.2$-$604207. During the outburst, detected in 2024 February by $\textit{eROSITA}$, the broadband persistent spectrum was well fitted by an absorbed Comptonization model. We detected 60 type I X-ray bursts in the $\textit{Insight}$-HXMT medium energy (ME) data, and 37 were also detected with the low-energy (LE) telescope. By superimposing the $\textit{Insight}$-HXMT/LE/ME/HE light curves of 37 bursts with similar profiles and intensities, we measured a deficit of X-rays in the 40$-$70 keV energy band. By analyzing the time-resolved X-ray burst spectra, we determine the mean ratio of persistent to burst flux of $\alpha=71\pm7$. We estimate the average hydrogen mass fraction in the fuel at ignition, as $\bar{X} = 0.342 \pm 0.033$, and constrain the burst fuel composition as $X_0\lesssim0.4$. We found that 14 out of 60 X-ray bursts exhibited photospheric expansion, and thus we estimated the distance to the source as $10.03\pm 0.71$ kpc. Combined with $\textit{IXPE}$ observations, the burst recurrence time were increasing from 1.55 to 8 hr as the local mass accretion rate decreasing, which can be described as $\Delta T_{\rm rec}\sim \dot{m}^{-0.91\pm0.02}$., Comment: 13 pages, 9 figures, submitted to ApJ, comments are welcome

We present a JWST MIRI/MRS spectrum of the inner disk of WISE J044634.16$-$262756.1B (hereafter J0446B), an old ($\sim$34 Myr) M4.5 star but with hints of ongoing accretion. The spectrum is molecule-rich and dominated by hydrocarbons. We detect 14 molecular species (H$_2$, CH$_3$, CH$_4$, C$_2$H$_2$, $^{13}$CCH$_2$, C$_2$H$_4$, C$_2$H$_6$, C$_3$H$_4$, C$_4$H$_2$, C$_6$H$_6$, HCN, HC$_3$N, CO$_2$ and $^{13}$CO$_2$) and 2 atomic lines ([Ne II] and [Ar II]), all observed for the first time in a disk at this age. The detection of spatially unresolved H$_2$ and Ne gas strongly supports that J0446B hosts a long-lived primordial disk, rather than a debris disk. The marginal H$_2$O detection and the high C$_2$H$_2$/CO$_2$ column density ratio indicate that the inner disk of J0446B has a very carbon-rich chemistry, with a gas-phase C/O ratio $\gtrsim$2, consistent with what have been found in most primordial disks around similarly low-mass stars. In the absence of significant outer disk dust substructures, inner disks are expected to first become water-rich due to the rapid inward drift of icy pebbles, and evolve into carbon-rich as outer disk gas flows inward on longer timescales. The faint millimeter emission in such low-mass star disks implies that they may have depleted their outer icy pebble reservoir early and already passed the water-rich phase. Models with pebble drift and volatile transport suggest that maintaining a carbon-rich chemistry for tens of Myr likely requires a slowly evolving disk with $\alpha-$viscosity $\lesssim10^{-4}$. This study represents the first detailed characterization of disk gas at $\sim$30 Myr, strongly motivating further studies into the final stages of disk evolution., Comment: Accepted by ApJL. 8 figures in the main text

Using toric modifications and some compatibility we compute the local $p$-adic zeta function of a plane curve singularity. Thanks to the compatibility, we can work over the analytic change of variables formula for $p$-adic integrals, hence avoid adapting to the algebraic setting and Denef's formula., Comment: 23 pages

We studied the spectral energy distribution (SED) of 22 known AM~CVns with orbital periods ($P_{orb}$) larger than 35~min using multiwavelength public photometric data to estimate the effective temperature of the accreting white dwarf. We find an infrared (IR) excess in all systems when compared to a single blackbody, both when the disk should be extended and when it should be truncated by the accretor's magnetic field. This suggests a dominant contribution from the donor to the IR flux. When fitting two blackbodies, the temperature of the hot component decreases with $P_{orb}$, as expected by evolutionary models. Temperatures for systems with $35

Radiogenic neutrons emitted by detector materials are one of the most challenging backgrounds for the direct search of dark matter in the form of weakly interacting massive particles (WIMPs). To mitigate this background, the XENONnT experiment is equipped with a novel gadolinium-doped water Cherenkov detector, which encloses the xenon dual-phase time projection chamber (TPC). The neutron veto (NV) tags neutrons via their capture on gadolinium or hydrogen, which release $\gamma$-rays that are subsequently detected as Cherenkov light. In this work, we present the key features and the first results of the XENONnT NV when operated with demineralized water in the initial phase of the experiment. Its efficiency for detecting neutrons is $(82\pm 1)\,\%$, the highest neutron detection efficiency achieved in a water Cherenkov detector. This enables a high efficiency of $(53\pm 3)\,\%$ for the tagging of WIMP-like neutron signals, inside a tagging time window of $250\,\mathrm{\mu s}$ between TPC and NV, leading to a livetime loss of $1.6\,\%$ during the first science run of XENONnT.

Large language models (LLMs) are notoriously memory-intensive during training, particularly with the popular AdamW optimizer. This memory burden necessitates using more or higher-end GPUs or reducing batch sizes, limiting training scalability and throughput. To address this, various memory-efficient optimizers have been proposed to reduce optimizer memory usage. However, they face critical challenges: (i) reliance on costly SVD operations; (ii) significant performance trade-offs compared to AdamW; and (iii) still substantial optimizer memory overhead to maintain competitive performance. In this work, we identify that AdamW's learning rate adaptation rule can be effectively coarsened as a structured learning rate update. Based on this insight, we propose Approximated Gradient Scaling for Memory-Efficient LLM Optimization (APOLLO), which approximates learning rate scaling using an auxiliary low-rank optimizer state based on pure random projection. This structured learning rate update rule makes APOLLO highly tolerant to further memory reductions while delivering comparable pre-training performance. Even its rank-1 variant, APOLLO-Mini, achieves superior pre-training performance compared to AdamW with SGD-level memory costs. Extensive experiments demonstrate that the APOLLO series performs on-par with or better than AdamW, while achieving greater memory savings by nearly eliminating the optimization states of AdamW. These savings provide significant system-level benefits: (1) Enhanced Throughput: 3x throughput on an 8xA100-80GB setup compared to AdamW by supporting 4x larger batch sizes. (2) Improved Model Scalability: Pre-training LLaMA-13B with naive DDP on A100-80GB GPUs without system-level optimizations. (3) Low-End GPU Friendly Pre-training: Pre-training LLaMA-7B on a single GPU using less than 12 GB of memory with weight quantization., Comment: Preprint

Collaboration is a cornerstone of society. In the real world, human teammates make use of multi-sensory data to tackle challenging tasks in ever-changing environments. It is essential for embodied agents collaborating in visually-rich environments replete with dynamic interactions to understand multi-modal observations and task specifications. To evaluate the performance of generalizable multi-modal collaborative agents, we present TeamCraft, a multi-modal multi-agent benchmark built on top of the open-world video game Minecraft. The benchmark features 55,000 task variants specified by multi-modal prompts, procedurally-generated expert demonstrations for imitation learning, and carefully designed protocols to evaluate model generalization capabilities. We also perform extensive analyses to better understand the limitations and strengths of existing approaches. Our results indicate that existing models continue to face significant challenges in generalizing to novel goals, scenes, and unseen numbers of agents. These findings underscore the need for further research in this area. The TeamCraft platform and dataset are publicly available at https://github.com/teamcraft-bench/teamcraft.

Recent advances in large language models (LLMs) have shown significant promise, yet their evaluation raises concerns, particularly regarding data contamination due to the lack of access to proprietary training data. To address this issue, we present C$^2$LEVA, a comprehensive bilingual benchmark featuring systematic contamination prevention. C$^2$LEVA firstly offers a holistic evaluation encompassing 22 tasks, each targeting a specific application or ability of LLMs, and secondly a trustworthy assessment due to our contamination-free tasks, ensured by a systematic contamination prevention strategy that fully automates test data renewal and enforces data protection during benchmark data release. Our large-scale evaluation of 15 open-source and proprietary models demonstrates the effectiveness of C$^2$LEVA.

Memecoins, driven by social media engagement and cultural narratives, have rapidly grown within the Web3 ecosystem. Unlike traditional cryptocurrencies, they are shaped by humor, memes, and community sentiment. This paper introduces the Coin-Meme dataset, an open-source collection of visual, textual, community, and financial data from the Pump.fun platform on the Solana blockchain. We also propose a multimodal framework to analyze memecoins, uncovering patterns in cultural themes, community interaction, and financial behavior. Through clustering, sentiment analysis, and word cloud visualizations, we identify distinct thematic groups centered on humor, animals, and political satire. Additionally, we provide financial insights by analyzing metrics such as Market Entry Time and Market Capitalization, offering a comprehensive view of memecoins as both cultural artifacts and financial instruments within Web3. The Coin-Meme dataset is publicly available at https://github.com/hwlongCUHK/Coin-Meme.git., Comment: 4 pages, 2 figures, conference

Simulating and controlling physical systems described by partial differential equations (PDEs) are crucial tasks across science and engineering. Recently, diffusion generative models have emerged as a competitive class of methods for these tasks due to their ability to capture long-term dependencies and model high-dimensional states. However, diffusion models typically struggle with handling system states with abrupt changes and generalizing to higher resolutions. In this work, we propose Wavelet Diffusion Neural Operator (WDNO), a novel PDE simulation and control framework that enhances the handling of these complexities. WDNO comprises two key innovations. Firstly, WDNO performs diffusion-based generative modeling in the wavelet domain for the entire trajectory to handle abrupt changes and long-term dependencies effectively. Secondly, to address the issue of poor generalization across different resolutions, which is one of the fundamental tasks in modeling physical systems, we introduce multi-resolution training. We validate WDNO on five physical systems, including 1D advection equation, three challenging physical systems with abrupt changes (1D Burgers' equation, 1D compressible Navier-Stokes equation and 2D incompressible fluid), and a real-world dataset ERA5, which demonstrates superior performance on both simulation and control tasks over state-of-the-art methods, with significant improvements in long-term and detail prediction accuracy. Remarkably, in the challenging context of the 2D high-dimensional and indirect control task aimed at reducing smoke leakage, WDNO reduces the leakage by 33.2% compared to the second-best baseline.

Markovian projections arise in problems where we aim to mimic the one-dimensional marginal laws of an It\^o semimartingale by using another It\^o process with simpler dynamics. In applications, Markovian projections are useful in calibrating jump-diffusion models with both local and stochastic features, leading to the study of the inversion problems. In this paper, we invert the Markovian projections for pure jump processes, which can be used to construct calibrated local stochastic intensity (LSI) models for credit risk applications. Such models are jump process analogues of the notoriously hard to construct local stochastic volatility (LSV) models used in equity modeling., Comment: 30 pages