Search

3D Gaussian Splatting (3DGS) has become a crucial method for acquiring 3D assets. To protect the copyright of these assets, digital watermarking techniques can be applied to embed ownership information discreetly within 3DGS models. However, existing watermarking methods for meshes, point clouds, and implicit radiance fields cannot be directly applied to 3DGS models, as 3DGS models use explicit 3D Gaussians with distinct structures and do not rely on neural networks. Naively embedding the watermark on a pre-trained 3DGS can cause obvious distortion in rendered images. In our work, we propose an uncertainty-based method that constrains the perturbation of model parameters to achieve invisible watermarking for 3DGS. At the message decoding stage, the copyright messages can be reliably extracted from both 3D Gaussians and 2D rendered images even under various forms of 3D and 2D distortions. We conduct extensive experiments on the Blender, LLFF and MipNeRF-360 datasets to validate the effectiveness of our proposed method, demonstrating state-of-the-art performance on both message decoding accuracy and view synthesis quality.

We present the results of a search for gravitational-wave transients associated with core-collapse supernova SN 2023ixf, which was observed in the galaxy Messier 101 via optical emission on 2023 May 19th, during the LIGO-Virgo-KAGRA 15th Engineering Run. We define a five-day on-source window during which an accompanying gravitational-wave signal may have occurred. No gravitational waves have been identified in data when at least two gravitational-wave observatories were operating, which covered $\sim 14\%$ of this five-day window. We report the search detection efficiency for various possible gravitational-wave emission models. Considering the distance to M101 (6.7 Mpc), we derive constraints on the gravitational-wave emission mechanism of core-collapse supernovae across a broad frequency spectrum, ranging from 50 Hz to 2 kHz where we assume the GW emission occurred when coincident data are available in the on-source window. Considering an ellipsoid model for a rotating proto-neutron star, our search is sensitive to gravitational-wave energy $1 \times 10^{-5} M_{\odot} c^2$ and luminosity $4 \times 10^{-5} M_{\odot} c^2/\text{s}$ for a source emitting at 50 Hz. These constraints are around an order of magnitude more stringent than those obtained so far with gravitational-wave data. The constraint on the ellipticity of the proto-neutron star that is formed is as low as $1.04$, at frequencies above $1200$ Hz, surpassing results from SN 2019ejj., Comment: Main paper: 6 pages, 4 figures and 1 table. Total with appendices: 20 pages, 4 figures, and 1 table

The magnetar SGR 1935+2154 is the only known Galactic source of fast radio bursts (FRBs). FRBs from SGR 1935+2154 were first detected by CHIME/FRB and STARE2 in 2020 April, after the conclusion of the LIGO, Virgo, and KAGRA Collaborations' O3 observing run. Here we analyze four periods of gravitational wave (GW) data from the GEO600 detector coincident with four periods of FRB activity detected by CHIME/FRB, as well as X-ray glitches and X-ray bursts detected by NICER and NuSTAR close to the time of one of the FRBs. We do not detect any significant GW emission from any of the events. Instead, using a short-duration GW search (for bursts $\leq$ 1 s) we derive 50\% (90\%) upper limits of $10^{48}$ ($10^{49}$) erg for GWs at 300 Hz and $10^{49}$ ($10^{50}$) erg at 2 kHz, and constrain the GW-to-radio energy ratio to $\leq 10^{14} - 10^{16}$. We also derive upper limits from a long-duration search for bursts with durations between 1 and 10 s. These represent the strictest upper limits on concurrent GW emission from FRBs., Comment: 15 pages of text including references, 4 figures, 5 tables

Single-view 3D reconstruction methods like Triplane Gaussian Splatting (TGS) have enabled high-quality 3D model generation from just a single image input within seconds. However, this capability raises concerns about potential misuse, where malicious users could exploit TGS to create unauthorized 3D models from copyrighted images. To prevent such infringement, we propose a novel image protection approach that embeds invisible geometry perturbations, termed "geometry cloaks", into images before supplying them to TGS. These carefully crafted perturbations encode a customized message that is revealed when TGS attempts 3D reconstructions of the cloaked image. Unlike conventional adversarial attacks that simply degrade output quality, our method forces TGS to fail the 3D reconstruction in a specific way - by generating an identifiable customized pattern that acts as a watermark. This watermark allows copyright holders to assert ownership over any attempted 3D reconstructions made from their protected images. Extensive experiments have verified the effectiveness of our geometry cloak. Our project is available at https://qsong2001.github.io/geometry_cloak., Comment: Accepted by NeurIPS 2024

Highly performant large-scale pre-trained models promise to also provide a valuable foundation for learning specialized tasks, by fine-tuning the model to the desired task. By starting from a good general-purpose model, the goal is to achieve both specialization in the target task and maintain robustness. To assess the robustness of models to out-of-distribution samples after fine-tuning on downstream datasets, we introduce a new robust fine-tuning benchmark, ImageNet-RIB (Robustness Inheritance Benchmark). The benchmark consists of a set of related but distinct specialized (downstream) tasks; pre-trained models are fine-tuned on one task in the set and their robustness is assessed on the rest, iterating across all tasks for fine-tuning and assessment. We find that the continual learning methods, EWC and LwF maintain robustness after fine-tuning though fine-tuning generally does reduce performance on generalization to related downstream tasks across models. Not surprisingly, models pre-trained on large and rich datasets exhibit higher initial robustness across datasets and suffer more pronounced degradation during fine-tuning. The distance between the pre-training and downstream datasets, measured by optimal transport, predicts this performance degradation on the pre-training dataset. However, counterintuitively, model robustness after fine-tuning on related downstream tasks is the worst when the pre-training dataset is the richest and the most diverse. This suggests that starting with the strongest foundation model is not necessarily the best approach for performance on specialist tasks. The benchmark thus offers key insights for developing more resilient fine-tuning strategies and building robust machine learning models. https://jd730.github.io/projects/ImageNet-RIB

Long-tail learning has garnered widespread attention and achieved significant progress in recent times. However, even with pre-trained prior knowledge, models still exhibit weaker generalization performance on tail classes. The promising Sharpness-Aware Minimization (SAM) can effectively improve the generalization capability of models by seeking out flat minima in the loss landscape, which, however, comes at the cost of doubling the computational time. Since the update rule of SAM necessitates two consecutive (non-parallelizable) forward and backpropagation at each step. To address this issue, we propose a novel method called Random SAM prompt tuning (RSAM-PT) to improve the model generalization, requiring only one-step gradient computation at each step. Specifically, we search for the gradient descent direction within a random neighborhood of the parameters during each gradient update. To amplify the impact of tail-class samples and avoid overfitting, we employ the deferred re-weight scheme to increase the significance of tail-class samples. The classification accuracy of long-tailed data can be significantly improved by the proposed RSAM-PT, particularly for tail classes. RSAM-PT achieves the state-of-the-art performance of 90.3\%, 76.5\%, and 50.1\% on benchmark datasets CIFAR100-LT (IF 100), iNaturalist 2018, and Places-LT, respectively. The source code is temporarily available at https://github.com/Keke921/GNM-PT., Comment: NeurIPS 2024

One-shot Federated Learning (OFL) significantly reduces communication costs in FL by aggregating trained models only once. However, the performance of advanced OFL methods is far behind the normal FL. In this work, we provide a causal view to find that this performance drop of OFL methods comes from the isolation problem, which means that local isolatedly trained models in OFL may easily fit to spurious correlations due to the data heterogeneity. From the causal perspective, we observe that the spurious fitting can be alleviated by augmenting intermediate features from other clients. Built upon our observation, we propose a novel learning approach to endow OFL with superb performance and low communication and storage costs, termed as FuseFL. Specifically, FuseFL decomposes neural networks into several blocks, and progressively trains and fuses each block following a bottom-up manner for feature augmentation, introducing no additional communication costs. Comprehensive experiments demonstrate that FuseFL outperforms existing OFL and ensemble FL by a significant margin. We conduct comprehensive experiments to show that FuseFL supports high scalability of clients, heterogeneous model training, and low memory costs. Our work is the first attempt using causality to analyze and alleviate data heterogeneity of OFL.

We demonstrate that architectures which traditionally are considered to be ill-suited for a task can be trained using inductive biases from another architecture. Networks are considered untrainable when they overfit, underfit, or converge to poor results even when tuning their hyperparameters. For example, plain fully connected networks overfit on object recognition while deep convolutional networks without residual connections underfit. The traditional answer is to change the architecture to impose some inductive bias, although what that bias is remains unknown. We introduce guidance, where a guide network guides a target network using a neural distance function. The target is optimized to perform well and to match its internal representations, layer-by-layer, to those of the guide; the guide is unchanged. If the guide is trained, this transfers over part of the architectural prior and knowledge of the guide to the target. If the guide is untrained, this transfers over only part of the architectural prior of the guide. In this manner, we can investigate what kinds of priors different architectures place on untrainable networks such as fully connected networks. We demonstrate that this method overcomes the immediate overfitting of fully connected networks on vision tasks, makes plain CNNs competitive to ResNets, closes much of the gap between plain vanilla RNNs and Transformers, and can even help Transformers learn tasks which RNNs can perform more easily. We also discover evidence that better initializations of fully connected networks likely exist to avoid overfitting. Our method provides a mathematical tool to investigate priors and architectures, and in the long term, may demystify the dark art of architecture creation, even perhaps turning architectures into a continuous optimizable parameter of the network., Comment: Under Review; 24 pages, 9 figures; Project page and code is at https://untrainable-networks.github.io/

Cross-lingual summarization (CLS) aims to generate a summary for the source text in a different target language. Currently, instruction-tuned large language models (LLMs) excel at various English tasks. However, unlike languages such as English, Chinese or Spanish, for those relatively low-resource languages with limited usage or data, recent studies have shown that LLMs' performance on CLS tasks remains unsatisfactory even with few-shot settings. This raises the question: Are LLMs capable of handling cross-lingual summarization tasks for low-resource languages? To resolve this question, we fully explore the potential of large language models on cross-lingual summarization task for low-resource languages through our four-step zero-shot method: Summarization, Improvement, Translation and Refinement (SITR) with correspondingly designed prompts. We test our proposed method with multiple LLMs on two well-known cross-lingual summarization datasets with various low-resource target languages. The results show that: i) GPT-3.5 and GPT-4 significantly and consistently outperform other baselines when using our zero-shot SITR methods. ii) By employing our proposed method, we unlock the potential of LLMs, enabling them to effectively handle cross-lingual summarization tasks for relatively low-resource languages.

Momentum-Aided Prompt Optimization (MAPO) enhances the efficiency and efficacy of prompt optimization for Large Language Models (LLMs). Building on ProTeGi, MAPO uses positive natural language "gradients" and a momentum-based extension to refine prompts effectively. By tracking gradient history, MAPO avoids local minima and oscillations. It also utilizes beam search and an Upper Confidence Bound (UCB) algorithm for balanced candidate expansion and selection. Benchmark testing shows that MAPO achieves faster convergence time with fewer API calls and higher F1 scores than ProTeGi, proving it as a robust and scalable solution for automated prompt engineering in LLMs.

Recently, prompt learning has emerged as the state-of-the-art (SOTA) for fair text-to-image (T2I) generation. Specifically, this approach leverages readily available reference images to learn inclusive prompts for each target Sensitive Attribute (tSA), allowing for fair image generation. In this work, we first reveal that this prompt learning-based approach results in degraded sample quality. Our analysis shows that the approach's training objective -- which aims to align the embedding differences of learned prompts and reference images -- could be sub-optimal, resulting in distortion of the learned prompts and degraded generated images. To further substantiate this claim, as our major contribution, we deep dive into the denoising subnetwork of the T2I model to track down the effect of these learned prompts by analyzing the cross-attention maps. In our analysis, we propose a novel prompt switching analysis: I2H and H2I. Furthermore, we propose new quantitative characterization of cross-attention maps. Our analysis reveals abnormalities in the early denoising steps, perpetuating improper global structure that results in degradation in the generated samples. Building on insights from our analysis, we propose two ideas: (i) Prompt Queuing and (ii) Attention Amplification to address the quality issue. Extensive experimental results on a wide range of tSAs show that our proposed method outperforms SOTA approach's image generation quality, while achieving competitive fairness. More resources at FairQueue Project site: https://sutd-visual-computing-group.github.io/FairQueue, Comment: Accepted in NeurIPS24

Integrating multi-modal clinical data, such as electronic health records (EHR) and chest X-ray images (CXR), is particularly beneficial for clinical prediction tasks. However, in a temporal setting, multi-modal data are often inherently asynchronous. EHR can be continuously collected but CXR is generally taken with a much longer interval due to its high cost and radiation dose. When clinical prediction is needed, the last available CXR image might have been outdated, leading to suboptimal predictions. To address this challenge, we propose DDL-CXR, a method that dynamically generates an up-to-date latent representation of the individualized CXR images. Our approach leverages latent diffusion models for patient-specific generation strategically conditioned on a previous CXR image and EHR time series, providing information regarding anatomical structures and disease progressions, respectively. In this way, the interaction across modalities could be better captured by the latent CXR generation process, ultimately improving the prediction performance. Experiments using MIMIC datasets show that the proposed model could effectively address asynchronicity in multimodal fusion and consistently outperform existing methods., Comment: Accepted by NeurIPS-24

Background: The Cox model and its extensions assuming proportional hazards is widely used to estimate vaccine efficacy (VE). In the typical situation that VE wanes over time, the VE estimates are not only sensitive to study duration and timing of vaccine delivery in relation to disease seasonality but also biased in the presence of sample attrition. Furthermore, estimates of vaccine impact such as number of cases averted (NCA) are sensitive to background disease incidence and timing of vaccine delivery. Comparison of the estimates between trials with different features can be misleading. Methods: We propose estimation of VE as a function of time in the Cox model framework, using the area under the VE curve as a summary measure of VE, and extension of the method to estimate vaccine impact. We use simulations and re-analysis of a RTS,S/AS01 malaria vaccine trial dataset to demonstrate their properties and applications. Results: Simulation under scenarios with different trial duration, magnitude of sample attrition and timing of vaccine delivery, all assuming vaccine protection wanes over time, demonstrated the problems of conventional methods assuming proportional hazard, robustness and unbiasedness of the proposed methods, and comparability of the proposed estimates of vaccine efficacy and impact across trials with different features. Furthermore, the proposed NCA estimators are informative in determining the optimal vaccine delivery strategy in regions with highly seasonal disease transmission. Conclusions: The proposed method based on estimation of vaccine efficacy trajectory provides a robust, unbiased, and flexible approach to evaluate vaccine effects., Comment: 3667 words, 2 figures, 4 tables, 3 supplemental materials

We introduce Controller-Embedded Language Model Interactions (CELI), a framework that integrates control logic directly within language model (LM) prompts, facilitating complex, multi-stage task execution. CELI addresses limitations of existing prompt engineering and workflow optimization techniques by embedding control logic directly within the operational context of language models, enabling dynamic adaptation to evolving task requirements. Our framework transfers control from the traditional programming execution environment to the LMs, allowing them to autonomously manage computational workflows while maintaining seamless interaction with external systems and functions. CELI supports arbitrary function calls with variable arguments, bridging the gap between LMs' adaptive reasoning capabilities and conventional software paradigms' structured control mechanisms. To evaluate CELI's versatility and effectiveness, we conducted case studies in two distinct domains: code generation (HumanEval benchmark) and multi-stage content generation (Wikipedia-style articles). The results demonstrate notable performance improvements across a range of domains. CELI achieved a 4.9 percentage point improvement over the best reported score of the baseline GPT-4 model on the HumanEval code generation benchmark. In multi-stage content generation, 94.4% of CELI-produced Wikipedia-style articles met or exceeded first draft quality when optimally configured, with 44.4% achieving high quality. These outcomes underscore CELI's potential for optimizing AI-driven workflows across diverse computational domains., Comment: 26 pages, 2 figures

The wolensing Python package offers a solution for gravitational wave lensing computations within the full wave-optics regime. This tool is primarily designed to calculate the gravitational lensing amplification factor including diffractive effects, an essential component for generating accurate lensed gravitational wave waveforms. These waveforms are integral to astrophysical and cosmological studies related to gravitational-wave lensing. Integrating with lensingGW (Pagano, Hannuksela, and Li 2020), wolensing provides solutions for image positions in the high-frequency regime where wave and geometrical optics converge. This functionality allows the amplification factor to be applicable across a wider frequency range. Another key feature of wolensing is its ability to plot time delay contours on the lens plane, offering researchers a visual tool to better understand the relationship between the lens system and the amplification factor. wolensing is compatible with various lens models in lenstronomy (Birrer et al. 2021). There are also built-in lens models including point mass, singular isothermal sphere (SIS),and nonsingular isothermal ellipsoid (NIE) with jax (Bradbury et al. 2018) supporting GPU computation. Users can accommodate different lens models in the code with jax. wolensing is available as an open-source package on PyPI and can be installed via pip.

Challenge sets such as the Winograd Schema Challenge (WSC) are used to benchmark systems' ability to resolve ambiguities in natural language. If one assumes as in existing work that solving a given challenge set is at least as difficult as solving some more general task, then high performance on the challenge set should indicate high performance on the general task overall. However, we show empirically that this assumption of difficulty does not always hold. In particular, we demonstrate that despite the strong performance of prompted language models (LMs) on the WSC and its variants, these same modeling techniques perform relatively poorly at resolving certain pronominal ambiguities attested in OntoNotes and related datasets that are perceived to be easier. Motivated by these findings, we propose a method for ensembling a prompted LM with a supervised, task-specific system that is overall more accurate at resolving pronominal coreference across datasets. Finally, we emphasize that datasets involving the same linguistic phenomenon draw on distinct, but overlapping, capabilities, and evaluating on any one dataset alone does not provide a complete picture of a system's overall capability., Comment: CoNLL 2024

We develop a systems approach based on an energy-landscape concept to differentiate interactions involving redox activities and conformational changes of proteins and nucleic acids interactions in multi-layered protein-DNA regulatory networks under light disturbance. Our approach is a data-driven modeling workflow using a physics-informed machine learning algorithm to train a non-linear mathematical model for interpreting gene expression dynamics and to lead discovery for protein regulators using redox proteome analysis. We distinguish light-responsive elements within central carbon metabolism pathways from independent variables like circadian time using the publicly available transcriptome datasets of Synechococcus elongatus over diel cycles responding to light perturbations. Our approach provides interpretable de novo models for elucidating events of reactions in complex regulatory pathways in response to stressful disturbance from the environment. We discovered protein regulators in response to light disturbance in the proteome analysis involving shifts in protein abundance as well as cysteine redox states under constant illumination and after two hours of darkness. We discovered significant shifts in cysteine redox states in regulatory proteins such as transcription sigma factors and metabolic enzymes in the oxidative pentose phosphate pathway and the Calvin-Benson cycle, while the changes in their protein abundance were minimal. These results indicate that regulatory dynamics in reductant generation link photo-induced electron transport pathways and redox metabolic pathways with circadian rhythms through fast redox-induced conformational changes or slow expression regulations across networks.

We investigate the recently reported $t\bar{t}$ excess by the CMS Collaboration within the framework of conventional Two-Higgs-Doublet Models (2HDMs). Considering all four types (I, II, X, and Y), we perform a comprehensive parameter space scan using the best-fit values for a pseudoscalar boson $A$: $M_A = 365$ GeV, $\Gamma_A/M_A = 2\%$, and $\tan\beta = 1.28$. Theoretical requirements and experimental constraints are systematically applied, including conditions from a bounded-below scalar potential, vacuum stability, unitarity, perturbativity, Flavor-Changing Neutral Currents (FCNCs), and direct searches at high-energy colliders. Our analysis shows that perturbativity imposes upper bounds of around 723 GeV on $M_{H^\pm}$ and $M_H$. FCNC constraints exclude all viable parameter space in Types II and Y, while a small region persists in Types I and X, but this region is ultimately ruled out by recent $t\bar{t} Z$ measurements by the ATLAS and CMS Collaborations at the LHC. We conclude that conventional 2HDMs alone cannot accommodate a pseudoscalar boson that explains the observed $t\bar{t}$ excess within viable parameter space. However, incorporating toponium effects in the background fit could potentially alter this conclusion., Comment: 16 pages with 4 figures

We introduce the pCI software package for high-precision atomic structure calculations. The standard method of calculation is based on the configuration interaction (CI) method to describe valence correlations, but can be extended to attain better accuracy by including core correlations via many-body perturbation theory (CI+MBPT) or the all-order (CI+all-order) method, as well as QED corrections via QEDMOD. The software package enables calculations of atomic properties, including energy levels, g-factors, hyperfine structure constants, multipole transition matrix elements, polarizabilities, and isotope shifts. It also features modern high-performance computing paradigms, including dynamic memory allocations and large-scale parallelization via the message-passing interface, to optimize and accelerate computations., Comment: 22 pages, 6 tables, 2 figures

Uncertainty quantification is crucial to account for the imperfect predictions of machine learning algorithms for high-impact applications. Conformal prediction (CP) is a powerful framework for uncertainty quantification that generates calibrated prediction intervals with valid coverage. In this work, we study how CP intervals are affected by bias - the systematic deviation of a prediction from ground truth values - a phenomenon prevalent in many real-world applications. We investigate the influence of bias on interval lengths of two different types of adjustments -- symmetric adjustments, the conventional method where both sides of the interval are adjusted equally, and asymmetric adjustments, a more flexible method where the interval can be adjusted unequally in positive or negative directions. We present theoretical and empirical analyses characterizing how symmetric and asymmetric adjustments impact the "tightness" of CP intervals for regression tasks. Specifically for absolute residual and quantile-based non-conformity scores, we prove: 1) the upper bound of symmetrically adjusted interval lengths increases by $2|b|$ where $b$ is a globally applied scalar value representing bias, 2) asymmetrically adjusted interval lengths are not affected by bias, and 3) conditions when asymmetrically adjusted interval lengths are guaranteed to be smaller than symmetric ones. Our analyses suggest that even if predictions exhibit significant drift from ground truth values, asymmetrically adjusted intervals are still able to maintain the same tightness and validity of intervals as if the drift had never happened, while symmetric ones significantly inflate the lengths. We demonstrate our theoretical results with two real-world prediction tasks: sparse-view computed tomography (CT) reconstruction and time-series weather forecasting. Our work paves the way for more bias-robust machine learning systems., Comment: 17 pages, 6 figures, code available at: https://github.com/matthewyccheung/conformal-metric

This paper provides a detailed analysis of the NeuroPiano dataset, which comprise 104 audio recordings of student piano performances accompanied with 2255 textual feedback and ratings given by professional pianists. We offer a statistical overview of the dataset, focusing on the standardization of annotations and inter-annotator agreement across 12 evaluative questions concerning performance quality. We also explore the predictive relationship between audio features and teacher ratings via machine learning, as well as annotations provided for text analysis of the responses.

Periorbital segmentation and distance prediction using deep learning allows for the objective quantification of disease state, treatment monitoring, and remote medicine. However, there are currently no reports of segmentation datasets for the purposes of training deep learning models with sub mm accuracy on the regions around the eyes. All images (n=2842) had the iris, sclera, lid, caruncle, and brow segmented by five trained annotators. Here, we validate this dataset through intra and intergrader reliability tests and show the utility of the data in training periorbital segmentation networks. All the annotations are publicly available for free download. Having access to segmentation datasets designed specifically for oculoplastic surgery will permit more rapid development of clinically useful segmentation networks which can be leveraged for periorbital distance prediction and disease classification. In addition to the annotations, we also provide an open-source toolkit for periorbital distance prediction from segmentation masks. The weights of all models have also been open-sourced and are publicly available for use by the community., Comment: 10 pages, 7 figures, 3 Tables

Periorbital distances and features around the eyes and lids hold valuable information for disease quantification and monitoring of surgical and medical intervention. These distances are commonly measured manually, a process that is both subjective and highly time-consuming. Here, we set out to developed three deep-learning methods for segmentation and periorbital distance prediction, and also evaluate the utility of periorbital distances for disease classification. The MAE of our deep learning predicted distances was less than or very close to the error observed between trained human annotators. We compared our models to the current state-of-the-art (SOTA) method for periorbital distance prediction and found that our methods outperformed SOTA on all of our datasets on all but one periorbital measurement. We also show that robust segmentation can be achieved on diseased eyes using models trained on open-source, healthy eyes, and that periorbital distances have can be used as high-quality features in downstream classification models. Leveraging segmentation networks as intermediary steps in classification has broad implications for increasing the generalizability of classification models in ophthalmic plastic and craniofacial surgery by avoiding the out-of-distribution problem observed in traditional convolutional neural networks., Comment: 20 pages, 10 figures, 9 tables

This paper is devoted to a Feynman-Kac formula for general nonlinear time-dependent Schr\"odinger equations with applications in numerical approximations. Our formulation integrates both the Fisk-Stratonovich and It\^o integrals within the framework of backward stochastic differential equations. Utilizing this Feynman-Kac representation, we propose a deep-learning-based approach for numerical approximation. Numerical experiments are performed to validate the accuracy and efficiency of our method, and a convergence analysis is provided to support the results., Comment: 25 pages, 4 figures

Decentralized diagnostic testing that is accurate, portable, quantitative, and capable of making multiple simultaneous measurements of different biomarkers at the point-of-need remains an important unmet need in the post-pandemic world. Resonator-based biosensors using silicon photonic integrated circuits are a promising technology to meet this need, as they can leverage (1) semiconductor manufacturing economies of scale, (2) exquisite optical sensitivity, and (3) the ability to integrate tens to hundreds of sensors on a millimeter-scale photonic chip. However, their application to decentralized testing has historically been limited by the expensive, bulky tunable lasers and alignment optics required for their readout. In this work, we introduce a segmented sensor architecture that addresses this important challenge by facilitating resonance-tracking readout using a fixed-wavelength laser. The architecture incorporates an in-resonator phase shifter modulated by CMOS drivers to periodically sweep and acquire the resonance peak shifts as well as a distinct high-sensitivity sensing region, maintaining high performance at a fraction of the cost and size. We show, for the first time, that fixed-wavelength sensor readout can offer similar performance to traditional tunable laser readout, demonstrating a system limit of detection of 6.1 x 10-5 RIU as well as immunoassay-based detection of the SARS-CoV-2 spike protein. We anticipate that this sensor architecture will open the door to a new data-rich class of portable, accurate, multiplexed diagnostics for decentralized testing.

Timely disease diagnosis is challenging due to increasing disease burdens and limited clinician availability. AI shows promise in diagnosis accuracy but faces real-world application issues due to insufficient validation in clinical workflows and diverse populations. This study addresses gaps in medical AI downstream accountability through a case study on age-related macular degeneration (AMD) diagnosis and severity classification. We designed and implemented an AI-assisted diagnostic workflow for AMD, comparing diagnostic performance with and without AI assistance among 24 clinicians from 12 institutions with real patient data sampled from the Age-Related Eye Disease Study (AREDS). Additionally, we demonstrated continual enhancement of an existing AI model by incorporating approximately 40,000 additional medical images (named AREDS2 dataset). The improved model was then systematically evaluated using both AREDS and AREDS2 test sets, as well as an external test set from Singapore. AI assistance markedly enhanced diagnostic accuracy and classification for 23 out of 24 clinicians, with the average F1-score increasing by 20% from 37.71 (Manual) to 45.52 (Manual + AI) (P-value < 0.0001), achieving an improvement of over 50% in some cases. In terms of efficiency, AI assistance reduced diagnostic times for 17 out of the 19 clinicians tracked, with time savings of up to 40%. Furthermore, a model equipped with continual learning showed robust performance across three independent datasets, recording a 29% increase in accuracy, and elevating the F1-score from 42 to 54 in the Singapore population.

Aerial grasping, particularly soft aerial grasping, holds significant promise for drone delivery and harvesting tasks. However, controlling UAV dynamics during aerial grasping presents considerable challenges. The increased mass during payload grasping adversely affects thrust prediction, while unpredictable environmental disturbances further complicate control efforts. In this study, our objective aims to enhance the control of the Soft Aerial Vehicle (SAV) during aerial grasping by incorporating a disturbance observer into a Nonlinear Model Predictive Control (NMPC) SAV controller. By integrating the disturbance observer into the NMPC SAV controller, we aim to compensate for dynamic model idealization and uncertainties arising from additional payloads and unpredictable disturbances. Our approach combines a disturbance observer-based NMPC with the SAV controller, effectively minimizing tracking errors and enabling precise aerial grasping along all three axes. The proposed SAV equipped with Disturbance Observer-based Nonlinear Model Predictive Control (DOMPC) demonstrates remarkable capabilities in handling both static and non-static payloads, leading to the successful grasping of various objects. Notably, our SAV achieves an impressive payload-to-weight ratio, surpassing previous investigations in the domain of soft grasping. Using the proposed soft aerial vehicle weighing 1.002 kg, we achieve a maximum payload of 337 g by grasping., Comment: 8 pages, 10 figures, submitted to IEEE Robotics Automation Letters

Diffusion models have demonstrated remarkable efficacy across various image-to-image tasks. In this research, we introduce Imagine yourself, a state-of-the-art model designed for personalized image generation. Unlike conventional tuning-based personalization techniques, Imagine yourself operates as a tuning-free model, enabling all users to leverage a shared framework without individualized adjustments. Moreover, previous work met challenges balancing identity preservation, following complex prompts and preserving good visual quality, resulting in models having strong copy-paste effect of the reference images. Thus, they can hardly generate images following prompts that require significant changes to the reference image, \eg, changing facial expression, head and body poses, and the diversity of the generated images is low. To address these limitations, our proposed method introduces 1) a new synthetic paired data generation mechanism to encourage image diversity, 2) a fully parallel attention architecture with three text encoders and a fully trainable vision encoder to improve the text faithfulness, and 3) a novel coarse-to-fine multi-stage finetuning methodology that gradually pushes the boundary of visual quality. Our study demonstrates that Imagine yourself surpasses the state-of-the-art personalization model, exhibiting superior capabilities in identity preservation, visual quality, and text alignment. This model establishes a robust foundation for various personalization applications. Human evaluation results validate the model's SOTA superiority across all aspects (identity preservation, text faithfulness, and visual appeal) compared to the previous personalization models.

In recent years, spatial computing Virtual Reality (VR) has emerged as a transformative technology, offering users immersive and interactive experiences across diversified virtual environments. Users can interact with VR apps through interactable GUI elements (IGEs) on the stereoscopic three-dimensional (3D) graphical user interface (GUI). The accurate recognition of these IGEs is instrumental, serving as the foundation of many software engineering tasks, including automated testing and effective GUI search. The most recent IGE detection approaches for 2D mobile apps typically train a supervised object detection model based on a large-scale manually-labeled GUI dataset, usually with a pre-defined set of clickable GUI element categories like buttons and spinners. Such approaches can hardly be applied to IGE detection in VR apps, due to a multitude of challenges including complexities posed by open-vocabulary and heterogeneous IGE categories, intricacies of context-sensitive interactability, and the necessities of precise spatial perception and visual-semantic alignment for accurate IGE detection results. Thus, it is necessary to embark on the IGE research tailored to VR apps. In this paper, we propose the first zero-shot cOntext-sensitive inteRactable GUI ElemeNT dEtection framework for virtual Reality apps, named Orienter. By imitating human behaviors, Orienter observes and understands the semantic contexts of VR app scenes first, before performing the detection. The detection process is iterated within a feedback-directed validation and reflection loop. Specifically, Orienter contains three components, including (1) Semantic context comprehension, (2) Reflection-directed IGE candidate detection, and (3) Context-sensitive interactability classification. Extensive experiments demonstrate that Orienter is more effective than the state-of-the-art GUI element detection approaches.

The sum of the absolute values of the Fourier coefficients of a function $f:\mathbb{F}_2^n \to \mathbb{R}$ is called the spectral norm of $f$. Green and Sanders' quantitative version of Cohen's idempotent theorem states that if the spectral norm of $f:\mathbb{F}_2^n \to \{0,1\}$ is at most $M$, then the support of $f$ belongs to the ring of sets generated by at most $\ell(M)$ cosets, where $\ell(M)$ is a constant that only depends on $M$. We prove that the above statement can be generalized to \emph{approximate} spectral norms if and only if the support of $f$ and its complement satisfy a certain arithmetic connectivity condition. In particular, our theorem provides a new proof of the quantitative Cohen's theorem for $\mathbb{F}_2^n$.

As the demand for computer science (CS) skills grows, mastering foundational concepts is crucial yet challenging for novice learners. To address this challenge, we present KoroT-3E, an AI-based system that creates personalized musical mnemonics to enhance both memory retention and understanding of concepts in CS. KoroT-3E enables users to transform complex concepts into memorable lyrics and compose melodies that suit their musical preferences. We conducted semi-structured interviews (n=12) to investigate why novice learners find it challenging to memorize and understand CS concepts. The findings, combined with constructivist learning theory, established our initial design, which was then refined following consultations with CS education experts. An empirical experiment(n=36) showed that those using KoroT-3E (n=18) significantly outperformed the control group (n=18), with improved memory efficiency, increased motivation, and a positive learning experience. These findings demonstrate the effectiveness of integrating multimodal generative AI into CS education to create personalized and interactive learning experiences., Comment: 26 pages , 2 tables,7 figures, undereview

We report on spectroscopy of the low-lying electronic transitions in $^{224}$Ra$^+$. The ion's low charge to mass ratio and convenient wavelengths make $^{224}$Ra$^+$ a promising optical clock candidate. We measured the frequencies of the the $^2{S}_{1/2} \ $$\leftrightarrow$$\ ^2{P}_{1/2}$ cooling transition, the $^2{S}_{1/2}\ $$\leftrightarrow$$\ ^2{D}_{5/2}$ clock transition, the $^2{D}_{3/2} \ $$\leftrightarrow$$\ ^2{P}_{3/2}$ electric dipole transition, and the $^2{D}_{5/2} \ $$\leftrightarrow$$\ ^2{P}_{3/2}$ cleanout transition. From these measurements we calculate the frequencies of the $^2{D}_{3/2}\ $$\leftrightarrow$$\ ^2{P}_{1/2}$ repump transition, the $^2{S}_{1/2} \ $$\leftrightarrow$$\ ^2{D}_{3/2}$ electric quadrupole transition, and the $^2{S}_{1/2} \ $$\leftrightarrow$$\ ^2{P}_{3/2}$ electric dipole transition., Comment: 5 pages, 8 figures

Transition metal dichalcogenides (TMDCs) possess valley pseudospin, allowing photon spin to be coupled to electron spin and enabling initialization and readout of both classical and quantum information. Rapid valley-dephasing processes have impeded the development of scalable, high-performance valleytronic devices operating at room temperature. Here we demonstrate that a chiral resonant metasurface can enable room-temperature valley-selective emission, even with linearly polarized excitation. This platform provides circular eigen-polarization states with a high quality factor (Q-factor) and strong chiral near-field enhancement, resulting in unitary emission circular dichroism (i.e. single-handed circularly polarized emission). Our fabricated Si chiral metasurfaces exhibit chiral electromagnetic modes with Q-factors up to 450 at visible wavelengths, spectrally tuned to the exciton energy of MoSe2 monolayers. Using spatially- and spectrally-resolved mapping from temperatures of 100 K to 294 K, we demonstrate degrees of circular polarization (DOP) as high as 0.5 at room temperature. Reciprocal space mapping of the exciton emission reveals the chiral q-BIC localizes valley-selective emission in the vicinity of the photonic gamma-point. Photon-spin and time-resolved photoluminescence measurements show that the high DOP can be attributed to the significantly increased chiroptical local density of states provided by the metasurface, which enhances valley-specific radiative transition rates by a factor of approximately 13, with lifetimes as short as 189 ps. Our work could facilitate the development of compact chiral classical and quantum light sources and the creation of molecular chiral polaritons for quantum enantioselective synthesis.

Research in music understanding has extensively explored composition-level attributes such as key, genre, and instrumentation through advanced representations, leading to cross-modal applications using large language models. However, aspects of musical performance such as stylistic expression and technique remain underexplored, along with the potential of using large language models to enhance educational outcomes with customized feedback. To bridge this gap, we introduce LLaQo, a Large Language Query-based music coach that leverages audio language modeling to provide detailed and formative assessments of music performances. We also introduce instruction-tuned query-response datasets that cover a variety of performance dimensions from pitch accuracy to articulation, as well as contextual performance understanding (such as difficulty and performance techniques). Utilizing AudioMAE encoder and Vicuna-7b LLM backend, our model achieved state-of-the-art (SOTA) results in predicting teachers' performance ratings, as well as in identifying piece difficulty and playing techniques. Textual responses from LLaQo was moreover rated significantly higher compared to other baseline models in a user study using audio-text matching. Our proposed model can thus provide informative answers to open-ended questions related to musical performance from audio data.

Signed graphs are equipped with both positive and negative edge weights, encoding pairwise correlations as well as anti-correlations in data. A balanced signed graph has no cycles of odd number of negative edges. Laplacian of a balanced signed graph has eigenvectors that map simply to ones in a similarity-transformed positive graph Laplacian, thus enabling reuse of well-studied spectral filters designed for positive graphs. We propose a fast method to learn a balanced signed graph Laplacian directly from data. Specifically, for each node $i$, to determine its polarity $\beta_i \in \{-1,1\}$ and edge weights $\{w_{i,j}\}_{j=1}^N$, we extend a sparse inverse covariance formulation based on linear programming (LP) called CLIME, by adding linear constraints to enforce ``consistent" signs of edge weights $\{w_{i,j}\}_{j=1}^N$ with the polarities of connected nodes -- i.e., positive/negative edges connect nodes of same/opposing polarities. For each LP, we adapt projections on convex set (POCS) to determine a suitable CLIME parameter $\rho > 0$ that guarantees LP feasibility. We solve the resulting LP via an off-the-shelf LP solver in $\mathcal{O}(N^{2.055})$. Experiments on synthetic and real-world datasets show that our balanced graph learning method outperforms competing methods and enables the use of spectral filters and graph convolutional networks (GCNs) designed for positive graphs on signed graphs., Comment: 5 pages, 1 figure. Submitted to ICASSP 2025

An image denoiser can be used for a wide range of restoration problems via the Plug-and-Play (PnP) architecture. In this paper, we propose a general framework to build an interpretable graph-based deep denoiser (GDD) by unrolling a solution to a maximum a posteriori (MAP) problem equipped with a graph Laplacian regularizer (GLR) as signal prior. Leveraging a recent theorem showing that any (pseudo-)linear denoiser $\boldsymbol \Psi$, under mild conditions, can be mapped to a solution of a MAP denoising problem regularized using GLR, we first initialize a graph Laplacian matrix $\mathbf L$ via truncated Taylor Series Expansion (TSE) of $\boldsymbol \Psi^{-1}$. Then, we compute the MAP linear system solution by unrolling iterations of the conjugate gradient (CG) algorithm into a sequence of neural layers as a feed-forward network -- one that is amenable to parameter tuning. The resulting GDD network is "graph-interpretable", low in parameter count, and easy to initialize thanks to $\mathbf L$ derived from a known well-performing denoiser $\boldsymbol \Psi$. Experimental results show that GDD achieves competitive image denoising performance compared to competitors, but employing far fewer parameters, and is more robust to covariate shift.

One of the latest trends in generative Artificial Intelligence is tools that generate and analyze content in different modalities, such as text and images, and convert information from one to the other. From a conceptual point of view, it is interesting to study whether these modality changes incur information loss and to what extent. This is analogous to variants of the classical game telephone, where players alternate between describing images and creating drawings based on those descriptions leading to unexpected transformations of the original content. In the case of AI, modality changes can be applied recursively, starting from an image to extract a text that describes it; using the text to generate a second image, extracting a text that describes it, and so on. As this process is applied recursively, AI tools are generating content from one mode to use them to create content in another mode and so on. Ideally, the embeddings of all of them would remain close to those of the original content so that only small variations are observed in the generated content versus the original one. However, it may also be the case the distance to the original embeddings increases in each iteration leading to a divergence in the process and to content that is barely related to the original one. In this paper, we present the results of an empirical study on the impact of recursive modality changes using GPT-4o, a state-of-the-art AI multimodal tool, and DALL-E 3. The results show that the multimodality loop diverges from the initial image without converging to anything specific. We have observed differences depending on the type of initial image and the configuration of the models. These findings are particularly relevant due to the increasing use of these tools for content generation, reconstruction, and adaptation, and their potential implications for the content on the Internet of the future.

We present a complete analysis of Fermi Large Area Telescope (LAT) data of GRB 221009A, the brightest Gamma-Ray Burst (GRB) ever detected. The burst emission above 30 MeV detected by the LAT preceded by 1 s the low-energy (< 10 MeV) pulse that triggered the Fermi Gamma-Ray Burst Monitor (GBM), as has been observed in other GRBs. The prompt phase of GRB 221009A lasted a few hundred seconds. It was so bright that we identify a Bad Time Interval (BTI) of 64 seconds caused by the extremely high flux of hard X-rays and soft gamma rays, during which the event reconstruction efficiency was poor and the dead time fraction quite high. The late-time emission decayed as a power law, but the extrapolation of the late-time emission during the first 450 seconds suggests that the afterglow started during the prompt emission. We also found that high-energy events observed by the LAT are incompatible with synchrotron origin, and, during the prompt emission, are more likely related to an extra component identified as synchrotron self-Compton (SSC). A remarkable 400 GeV photon, detected by the LAT 33 ks after the GBM trigger and directionally consistent with the location of GRB 221009A, is hard to explain as a product of SSC or TeV electromagnetic cascades, and the process responsible for its origin is uncertain. Because of its proximity and energetic nature, GRB 221009A is an extremely rare event., Comment: 60 pages, 38 figures, 9 tables

We compute the differential cross section for direct quarkonium production in high-energy electron-nucleus collisions at small $x$. Our computation is performed within the nonrelativistic QCD factorization formalism that separates the calculation into short distance coefficients and long distance matrix elements that depend on the color and spin of the state. We obtain the short distance coefficients of the production of the heavy quark pair within the framework of the Color Glass Condensate effective field theory, which resums coherent multiple interactions of the heavy quark pair with the nucleus to all orders. Our results are expressed as the convolution of perturbatively calculable perturbative functions with multi-point light-like Wilson line correlators. In the correlation limit, we establish the correspondence between our CGC formulation with calculations employing the transverse momentum dependent (TMD) framework. We extend this correspondence by resumming kinematic power corrections within the improved TMD framework, which interpolates between the TMD formalism and $k_\perp$ factorization formalism. We present a detailed numerical analysis, focusing on $J/\psi$ production in the kinematics accessible at the future Electron-Ion Collider, highlighting the importance of genuine higher-order saturation contributions when the electron collides with a large nucleus., Comment: 31 pages, 8 figures

The physics case for quarkonium-production studies accessible at the US Electron Ion Collider is described., Comment: Latex, 84 pages. Review prepared for Progress in Particle and Nuclear Physics

Optimization over Stiefel manifolds has found wide applications in many scientific and engineering domains. Despite considerable research effort, high-dimensional optimization problems over Stiefel manifolds remain challenging, and the situation is exacerbated by nonsmooth objective functions. The purpose of this paper is to propose and study a novel coordinate-type algorithm for weakly convex (possibly nonsmooth) optimization problems over high-dimensional Stiefel manifolds, named randomized submanifold subgradient method (RSSM). Similar to coordinate-type algorithms in the Euclidean setting, RSSM exhibits low per-iteration cost and is suitable for high-dimensional problems. We prove that RSSM converges to the set of stationary points and attains $\varepsilon$-stationary points with respect to a natural stationarity measure in $\mathcal{O}(\varepsilon^{-4})$ iterations in both expectation and the almost-sure senses. To the best of our knowledge, these are the first convergence guarantees for coordinate-type algorithms to optimize nonconvex nonsmooth functions over Stiefel manifolds. An important technical tool in our convergence analysis is a new Riemannian subgradient inequality for weakly convex functions on proximally smooth matrix manifolds, which could be of independent interest.

Studying the magnetic field properties on the solar surface is crucial for understanding the solar and heliospheric activities, which in turn shape space weather in the solar system. Surface Flux Transport (SFT) modelling helps us to simulate and analyse the transport and evolution of magnetic flux on the solar surface, providing valuable insights into the mechanisms responsible for solar activity. In this work, we demonstrate the use of machine learning techniques in solving magnetic flux transport, making it accurate. We have developed a novel Physics-Informed Neural Networks (PINNs)-based model to study the evolution of Bipolar Magnetic Regions (BMRs) using SFT in one-dimensional azimuthally averaged and also in two-dimensions. We demonstrate the efficiency and computational feasibility of our PINNs-based model by comparing its performance and accuracy with that of a numerical model implemented using the Runge-Kutta Implicit-Explicit (RK-IMEX) scheme. The mesh-independent PINNs method can be used to reproduce the observed polar magnetic field with better flux conservation. This advancement is important for accurately reproducing observed polar magnetic fields, thereby providing insights into the strength of future solar cycles. This work paves the way for more efficient and accurate simulations of solar magnetic flux transport and showcases the applicability of PINNs in solving advection-diffusion equations with a particular focus on heliophysics., Comment: 21 pages, 10 figures

Skip connections are fundamental architecture designs for modern deep neural networks (DNNs) such as CNNs and ViTs. While they help improve model performance significantly, we identify a vulnerability associated with skip connections to Model Inversion (MI) attacks, a type of privacy attack that aims to reconstruct private training data through abusive exploitation of a model. In this paper, as a pioneer work to understand how DNN architectures affect MI, we study the impact of skip connections on MI. We make the following discoveries: 1) Skip connections reinforce MI attacks and compromise data privacy. 2) Skip connections in the last stage are the most critical to attack. 3) RepVGG, an approach to remove skip connections in the inference-time architectures, could not mitigate the vulnerability to MI attacks. 4) Based on our findings, we propose MI-resilient architecture designs for the first time. Without bells and whistles, we show in extensive experiments that our MI-resilient architectures can outperform state-of-the-art (SOTA) defense methods in MI robustness. Furthermore, our MI-resilient architectures are complementary to existing MI defense methods. Our project is available at https://Pillowkoh.github.io/projects/RoLSS/, Comment: Accepted by ECCV 2024

Large language models (LLMs) have demonstrated remarkable potential across numerous applications and have shown an emergent ability to tackle complex reasoning tasks, such as mathematical computations. However, even for the simplest arithmetic calculations, the intrinsic mechanisms behind LLMs remain mysterious, making it challenging to ensure reliability. In this work, we delve into uncovering a specific mechanism by which LLMs execute calculations. Through comprehensive experiments, we find that LLMs frequently involve a small fraction (< 5%) of attention heads, which play a pivotal role in focusing on operands and operators during calculation processes. Subsequently, the information from these operands is processed through multi-layer perceptrons (MLPs), progressively leading to the final solution. These pivotal heads/MLPs, though identified on a specific dataset, exhibit transferability across different datasets and even distinct tasks. This insight prompted us to investigate the potential benefits of selectively fine-tuning these essential heads/MLPs to boost the LLMs' computational performance. We empirically find that such precise tuning can yield notable enhancements on mathematical prowess, without compromising the performance on non-mathematical tasks. Our work serves as a preliminary exploration into the arithmetic calculation abilities inherent in LLMs, laying a solid foundation to reveal more intricate mathematical tasks., Comment: Accepted by ICML 2024 (oral)

Model Inversion (MI) is a type of privacy violation that focuses on reconstructing private training data through abusive exploitation of machine learning models. To defend against MI attacks, state-of-the-art (SOTA) MI defense methods rely on regularizations that conflict with the training loss, creating explicit tension between privacy protection and model utility. In this paper, we present a new method to defend against MI attacks. Our method takes a new perspective and focuses on training data. Our idea is based on a novel insight on Random Erasing (RE), which has been applied in the past as a data augmentation technique to improve the model accuracy under occlusion. In our work, we instead focus on applying RE for degrading MI attack accuracy. Our key insight is that MI attacks require significant amount of private training data information encoded inside the model in order to reconstruct high-dimensional private images. Therefore, we propose to apply RE to reduce private information presented to the model during training. We show that this can lead to substantial degradation in MI reconstruction quality and attack accuracy. Meanwhile, natural accuracy of the model is only moderately affected. Our method is very simple to implement and complementary to existing defense methods. Our extensive experiments of 23 setups demonstrate that our method can achieve SOTA performance in balancing privacy and utility of the models. The results consistently demonstrate the superiority of our method over existing defenses across different MI attacks, network architectures, and attack configurations., Comment: Under review. The first two authors contributed equally