Search

Large language models (LLMs), like ChatGPT, have shown that even trained with noisy prior data, they can generalize effectively to new tasks through in-context learning (ICL) and pre-training techniques. Motivated by this, we explore whether a similar approach can be applied to scientific foundation models (SFMs). Our methodology is structured as follows: (i) we collect low-cost physics-informed neural network (PINN)-based approximated prior data in the form of solutions to partial differential equations (PDEs) constructed through an arbitrary linear combination of mathematical dictionaries; (ii) we utilize Transformer architectures with self and cross-attention mechanisms to predict PDE solutions without knowledge of the governing equations in a zero-shot setting; (iii) we provide experimental evidence on the one-dimensional convection-diffusion-reaction equation, which demonstrate that pre-training remains robust even with approximated prior data, with only marginal impacts on test accuracy. Notably, this finding opens the path to pre-training SFMs with realistic, low-cost data instead of (or in conjunction with) numerical high-cost data. These results support the conjecture that SFMs can improve in a manner similar to LLMs, where fully cleaning the vast set of sentences crawled from the Internet is nearly impossible.

The thermal Hall effect has emerged as an ideal probe for investigating topological phenomena of charge-neutral excitation. Notably, it reveals crucial aspects of spin-lattice couplings that have been difficult to access for decades. However, the exchange striction mechanism from a lattice-induced change in exchange interaction has often been ignored in thermal Hall experiments. MnPS3 can offer a platform to study exchange striction on the thermal Hall effect due to its significant spin-lattice coupling and field-induced non-collinear spin configuration. Our thermal transport data show distinct temperature and field dependence of longitudinal thermal conductivity ($\kappa_{xx}$) and thermal Hall effect ($\kappa_{xy}$). By using detailed theoretical calculation, we found that the inclusion of the exchange striction is essential for a better description of both $\kappa_{xx}$ and $\kappa_{xy}$. Our result demonstrates the importance of the exchange striction mechanism for a complete understanding of the magnon-phonon-driven thermal Hall effect., Comment: 24 pages, 6 figures, accepted for publication in Phys. Rev. B

Real-life robot navigation involves more than just reaching a destination; it requires optimizing movements while addressing scenario-specific goals. An intuitive way for humans to express these goals is through abstract cues like verbal commands or rough sketches. Such human guidance may lack details or be noisy. Nonetheless, we expect robots to navigate as intended. For robots to interpret and execute these abstract instructions in line with human expectations, they must share a common understanding of basic navigation concepts with humans. To this end, we introduce CANVAS, a novel framework that combines visual and linguistic instructions for commonsense-aware navigation. Its success is driven by imitation learning, enabling the robot to learn from human navigation behavior. We present COMMAND, a comprehensive dataset with human-annotated navigation results, spanning over 48 hours and 219 km, designed to train commonsense-aware navigation systems in simulated environments. Our experiments show that CANVAS outperforms the strong rule-based system ROS NavStack across all environments, demonstrating superior performance with noisy instructions. Notably, in the orchard environment, where ROS NavStack records a 0% total success rate, CANVAS achieves a total success rate of 67%. CANVAS also closely aligns with human demonstrations and commonsense constraints, even in unseen environments. Furthermore, real-world deployment of CANVAS showcases impressive Sim2Real transfer with a total success rate of 69%, highlighting the potential of learning from human demonstrations in simulated environments for real-world applications., Comment: project page https://worv-ai.github.io/canvas

Multi-Q magnetic structures on triangular lattices, with their two-dimensional topological spin texture, have attracted significant interest. However, unambiguously confirming their formation by excluding the presence of three equally-populated single-Q domains remains challenging. In the metallic triangular lattice antiferromagnet Co1/3TaS2, two magnetic ground states have been suggested at different temperature ranges, with the low-temperature phase being a triple-Q structure corresponding to the highest-density Skyrmion lattice. Using inelastic neutron scattering (INS) and advanced spin dynamics simulations, we demonstrate a clear distinction in the excitation spectra between the single-Q and triple-Q phases of Co1/3TaS2 and, more generally, a triangular lattice. First, we refined the spin Hamiltonian by fitting the excitation spectra measured in its paramagnetic phase, allowing us to develop an unbiased model independent of magnetic ordering. Second, we observed that the two magnetically ordered phases in Co1/3TaS2 exhibit markedly different behaviors in their long-wavelength Goldstone modes. Our spin model, derived from the paramagnetic phase, confirms that these behaviors originate from the single-Q and triple-Q nature of the respective ordered phases, providing unequivocal evidence of the single-Q to triple-Q phase transition in Co1/3TaS2. Importantly, we propose that the observed contrast in the long-wavelength spin dynamics between the single-Q and triple-Q orderings is universal, offering a potentially unique way to distinguish a generic triple-Q ordering on a triangular lattice from its multi-domain single-Q counterparts. We describe its applicability with examples of similar hexagonal systems forming potential triple-Q orderings. (For the full abstract, please refer to the manuscript), Comment: 14 pages, 6 figures excluding Appendices

The Sejong dictionary dataset offers a valuable resource, providing extensive coverage of morphology, syntax, and semantic representation. This dataset can be utilized to explore linguistic information in greater depth. The labeled linguistic structures within this dataset form the basis for uncovering relationships between words and phrases and their associations with target verbs. This paper introduces a user-friendly web interface designed for the collection and consolidation of verb-related information, with a particular focus on subcategorization frames. Additionally, it outlines our efforts in mapping this information by aligning subcategorization frames with corresponding illustrative sentence examples. Furthermore, we provide a Python library that would simplify syntactic parsing and semantic role labeling. These tools are intended to assist individuals interested in harnessing the Sejong dictionary dataset to develop applications for Korean language processing., Comment: NAACL 2025 System Demonstrations (Submitted)

Most selective colleges implemented test-optional admissions during the pandemic, making college entrance exam scores optional for applicants. We draw on descriptive, two-way fixed effects, and event study methods to examine variation in test-optional implementation during the pandemic and how implementation relates to selectivity and enrollment. For "test-optional" colleges during the pandemic, we found substantial variation in policy type (e.g., test optional, test free) and whether the policy extended to all applicants and scholarship consideration. Findings suggest test-optional implementation related to increases in Black student enrollment, mostly at moderately selective colleges and when policies extended to all applicants and scholarships. At highly selective colleges, findings suggest test-optional implementation related to an increase in applications but not consistent gains in enrollment.

Letters of recommendation from school counselors are required to apply to many selective colleges and universities. Still, relatively little is known about how this non-standardized component may affect equity in admissions. We use cutting-edge natural language processing techniques to algorithmically analyze a national dataset of over 600,000 student applications and counselor recommendation letters submitted via the Common App platform. We examine how the length and topical content of letters (e.g., sentences about Personal Qualities, Athletics, Intellectual Promise, etc.) relate to student self-identified race/ethnicity, sex, and proxies for socioeconomic status. Paired with regression analyses, we explore whether demographic differences in letter characteristics persist when accounting for additional student, school, and counselor characteristics, as well as among letters written by the same counselor and among students with comparably competitive standardized test scores. We ultimately find large and noteworthy naïve differences in letter length and content across nearly all demographic groups, many in alignment with known inequities (e.g., many more sentences about Athletics among White and higher-SES students, longer letters and more sentences on Personal Qualities for private school students). However, these differences vary drastically based on the exact controls and comparison groups included -- demonstrating that the ultimate implications of these letter differences for equity hinges on exactly how and when letters are used in admissions processes (e.g., among which groups of students are they used to "break ties"?). Findings do not point to a clear recommendation whether institutions should keep or discard letter requirements, but reflect the importance of reading letters and overall applications in the context of structural opportunity. We discuss additional implications and possible recommendations for college access and admissions policy/practice.

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

How to best develop foundational models for time series forecasting remains an important open question. Tokenization is a crucial consideration in this effort: what is an effective discrete vocabulary for a real-valued sequential input? To address this question, we develop WaveToken, a wavelet-based tokenizer that allows models to learn complex representations directly in the space of time-localized frequencies. Our method first scales and decomposes the input time series, then thresholds and quantizes the wavelet coefficients, and finally pre-trains an autoregressive model to forecast coefficients for the forecast horizon. By decomposing coarse and fine structures in the inputs, wavelets provide an eloquent and compact language for time series forecasting that simplifies learning. Empirical results on a comprehensive benchmark, including 42 datasets for both in-domain and zero-shot settings, show that WaveToken: i) provides better accuracy than recently proposed foundation models for forecasting while using a much smaller vocabulary (1024 tokens), and performs on par or better than modern deep learning models trained specifically on each dataset; and ii) exhibits superior generalization capabilities, achieving the best average rank across all datasets for three complementary metrics. In addition, we show that our method can easily capture complex temporal patterns of practical relevance that are challenging for other recent pre-trained models, including trends, sparse spikes, and non-stationary time series with varying frequencies evolving over time., Comment: 25 pages, 15 figures

We analyzed the 7.92$\times 10^{11}$ cosmic-ray-induced muon events collected by the IceCube Neutrino Observatory from May 13, 2011, when the fully constructed experiment started to take data, to May 12, 2023. This dataset provides an up-to-date cosmic-ray arrival direction distribution in the Southern Hemisphere with unprecedented statistical accuracy covering more than a full period length of a solar cycle. Improvements in Monte Carlo event simulation and better handling of year-to-year differences in data processing significantly reduce systematic uncertainties below the level of statistical fluctuations compared to the previously published results. We confirm the observation of a change in the angular structure of the cosmic-ray anisotropy between 10 TeV and 1 PeV, more specifically in the 100-300 TeV energy range.

Most prediction models for crystal properties employ a unimodal perspective, with graph-based representations, overlooking important non-local information that affects crystal properties. Some recent studies explore the impact of integrating graph and textual information on crystal property predictions to provide the model with this "missing" information by concatenation of embeddings. However, such studies do not evaluate which type of textual information is actually beneficial. We concatenate graph representations with text representations derived from textual descriptions with varying levels of detail. These descriptions, generated using the Robocrystallographer package, encompass global (e.g., space group, crystal type), local (e.g., bond lengths, coordination environment), and semiglobal (e.g., connectivity, arrangements) information about the structures. Our approach investigates how augmenting graph-based information with various levels of textual detail influences the performance for predictions for shear modulus and bulk modulus. We demonstrate that while graph representations can capture local structural information, incorporating semiglobal textual information enhances model performance the most. Global information can support performance further in the presence of semiglobal information. Our findings suggest that the strategic inclusion of textual information can enhance property prediction, thereby advancing the design and discovery of advanced novel materials for battery electrodes, catalysts, etc., Comment: 8 pages, 1 figure

The thermal Sunyaev-Zel'dovich (tSZ) effect arises from inverse Compton scattering of low energy photons onto thermal electrons, proportional to the integrated electron pressure, and is usually observed from galaxy clusters. However, we can expect that the Epoch of Reionization (EoR) also contributes to this signal, but that contribution has not been previously evaluated. In this work we analyse a suite of fully-coupled radiation-hydrodynamics simulations based on RAMSES-CUDATON to calculate and study the tSZ signal from the Reionization Epoch. We construct lightcones of the electron pressure in the intergalactic medium for $6

Understanding the internal computations of large language models (LLMs) is crucial for aligning them with human values and preventing undesirable behaviors like toxic content generation. However, mechanistic interpretability is hindered by polysemanticity -- where individual neurons respond to multiple, unrelated concepts. While Sparse Autoencoders (SAEs) have attempted to disentangle these features through sparse dictionary learning, they have compromised LLM performance due to reliance on post-hoc reconstruction loss. To address this issue, we introduce Mixture of Monosemantic Experts for Transformers (Monet) architecture, which incorporates sparse dictionary learning directly into end-to-end Mixture-of-Experts pretraining. Our novel expert decomposition method enables scaling the expert count to 262,144 per layer while total parameters scale proportionally to the square root of the number of experts. Our analyses demonstrate mutual exclusivity of knowledge across experts and showcase the parametric knowledge encapsulated within individual experts. Moreover, Monet allows knowledge manipulation over domains, languages, and toxicity mitigation without degrading general performance. Our pursuit of transparent LLMs highlights the potential of scaling expert counts to enhance} mechanistic interpretability and directly resect the internal knowledge to fundamentally adjust} model behavior. The source code and pretrained checkpoints are available at https://github.com/dmis-lab/Monet.

Document comparison typically relies on optical character recognition (OCR) as its core technology. However, OCR requires the selection of appropriate language models for each document and the performance of multilingual or hybrid models remains limited. To overcome these challenges, we propose text change detection (TCD) using an image comparison model tailored for multilingual documents. Unlike OCR-based approaches, our method employs word-level text image-to-image comparison to detect changes. Our model generates bidirectional change segmentation maps between the source and target documents. To enhance performance without requiring explicit text alignment or scaling preprocessing, we employ correlations among multi-scale attention features. We also construct a benchmark dataset comprising actual printed and scanned word pairs in various languages to evaluate our model. We validate our approach using our benchmark dataset and public benchmarks Distorted Document Images and the LRDE Document Binarization Dataset. We compare our model against state-of-the-art semantic segmentation and change detection models, as well as to conventional OCR-based models., Comment: 15pages, 11figures 6tables, wacv2025 accepted

We propose a Lorentzian wormhole geometry characterized by a closed string massless sector with nontrivial $H$-flux and a scalar dilaton. In the string frame, the dilaton exhibits a negative kinetic term, enabling the existence of the wormhole. The geometry consists of three distinct regions. The middle region contains the throat, and its boundaries with the other two regions form non-Riemannian two-spheres, where a fundamental string becomes chiral, akin to a non-relativistic string. While point-particle geodesics are complete within each region and non-traversable across regions, strings perceive the geometry differently, allowing a chiral string to traverse freely., Comment: 6 pages with 2 figures

Photonics has unlocked the potential for energy-efficient acceleration of deep learning. Most approaches toward photonic deep learning have diligently reproduced traditional deep learning architectures using photonic platforms, separately implementing linear-optical matrix calculations and nonlinear activations via electro-optical conversion, optical nonlinearities, and signal-encoded materials. Here we propose a concept of nonlinear unitary photonic circuits to achieve the integration of linear and nonlinear expressivity essential for deep neural networks. We devise a building block for two-dimensional nonlinear unitary operations, featuring norm-preserving mappings with nonconservative inner products, which enables the construction of high-dimensional nonlinear unitary circuits. Using deep nonlinear unitary circuits, we demonstrate exponential growth in trajectory length and near-complete coverage of the output space, both of which are essential for deep learning. Along with neuroevolutionary learning examples for the regression of a nonconvex function, our results pave the way to photonic neural networks with highly expressive inference and stable training., Comment: 33 pages, 7 figures

Maritime environmental sensing requires overcoming challenges from complex conditions such as harsh weather, platform perturbations, large dynamic objects, and the requirement for long detection ranges. While cameras and LiDAR are commonly used in ground vehicle navigation, their applicability in maritime settings is limited by range constraints and hardware maintenance issues. Radar sensors, however, offer robust long-range detection capabilities and resilience to physical contamination from weather and saline conditions, making it a powerful sensor for maritime navigation. Among various radar types, X-band radar (e.g., marine radar) is widely employed for maritime vessel navigation, providing effective long-range detection essential for situational awareness and collision avoidance. Nevertheless, it exhibits limitations during berthing operations where close-range object detection is critical. To address this shortcoming, we incorporate W-band radar (e.g., Navtech imaging radar), which excels in detecting nearby objects with a higher update rate. We present a comprehensive maritime sensor dataset featuring multi-range detection capabilities. This dataset integrates short-range LiDAR data, medium-range W-band radar data, and long-range X-band radar data into a unified framework. Additionally, it includes object labels for oceanic object detection usage, derived from radar and stereo camera images. The dataset comprises seven sequences collected from diverse regions with varying levels of estimation difficulty, ranging from easy to challenging, and includes common locations suitable for global localization tasks. This dataset serves as a valuable resource for advancing research in place recognition, odometry estimation, SLAM, object detection, and dynamic object elimination within maritime environments. Dataset can be found in following link: https://sites.google.com/view/rpmmoana, Comment: 8 pages, 8 figures, 3 tables

In this short note, we address the identifiability issues inherent in the Degree-Corrected Stochastic Block Model (DCSBM). We provide a rigorous proof demonstrating that the parameters of the DCSBM are identifiable up to a scaling factor and a permutation of the community labels, under a mild condition.

The goal of this work was to compute the semantic similarity among publicly available health survey questions in order to facilitate the standardization of survey-based Person-Generated Health Data (PGHD). We compiled various health survey questions authored in both English and Korean from the NIH CDE Repository, PROMIS, Korean public health agencies, and academic publications. Questions were drawn from various health lifelog domains. A randomized question pairing scheme was used to generate a Semantic Text Similarity (STS) dataset consisting of 1758 question pairs. Similarity scores between each question pair were assigned by two human experts. The tagged dataset was then used to build three classifiers featuring: Bag-of-Words, SBERT with BERT-based embeddings, and SBRET with LaBSE embeddings. The algorithms were evaluated using traditional contingency statistics. Among the three algorithms, SBERT-LaBSE demonstrated the highest performance in assessing question similarity across both languages, achieving an Area Under the Receiver Operating Characteristic (ROC) and Precision-Recall Curves of over 0.99. Additionally, it proved effective in identifying cross-lingual semantic similarities.The SBERT-LaBSE algorithm excelled at aligning semantically equivalent sentences across both languages but encountered challenges in capturing subtle nuances and maintaining computational efficiency. Future research should focus on testing with larger multilingual datasets and on calibrating and normalizing scores across the health lifelog domains to improve consistency. This study introduces the SBERT-LaBSE algorithm for calculating semantic similarity across two languages, showing it outperforms BERT-based models and the Bag of Words approach, highlighting its potential to improve semantic interoperability of survey-based PGHD across language barriers.

The 2175{\AA} bump is a prominent absorption feature at ultraviolet (UV) wavelengths in dust extinction and attenuation curves. Understanding the relative strength of this feature is important for accurate dust corrections at both low- and high-redshift. This feature is postulated to arise from polycyclic aromatic hydrocarbon (PAH) dust grains; however, the carrier has not been definitively established. We present results on the correlation between the 2175{\AA} feature and PAH abundances in a spatially-resolved manner for 15 local galaxies in the PHANGS-JWST survey that have NUV and mid-IR imaging data from Swift/UVOT and JWST/MIRI, respectively. We find a moderate positive correlation between the 2175{\AA} feature strength and PAH abundance, albeit with large intrinsic scatter. However, most of this trend can be attributed to a stronger negative correlation of both quantities with SFR surface density and specific-SFR (proxies of ionising radiation). The latter trends are consistent with previous findings that both the 2175{\AA} carrier and PAHs are small grains that are easily destroyed by UV photons, although the proxy for PAH abundance could also be influenced by dust heating. When controlling for SFR surface density, we find weaker correlations between the 2175{\AA} feature and PAH abundances, disfavouring a direct link. However, analyses based on spectroscopic measurements of the 2175{\AA} feature and PAH features are required to verify our findings. No significant trends with gas-phase metallicity are found for the 2175{\AA} feature and PAHs, however the metallicity range of our sample is limited. We provide prescriptions for the strength of the 2175{\AA} feature and PAHs in local massive (metal-rich) galaxies with SFR surface density and specific-SFR, however the former should be used with caution since bump strengths measured from Swift/UVOT are expected to be underestimated., Comment: 26 pages, 19 figures, 4 tables. Accepted for publication in PASA

We report JWST NIRSpec/G395H observations of TOI-1685 b, a hot rocky super-Earth orbiting an M2.5V star, during a full orbit. We obtain transmission and emission spectra of the planet and characterize the properties of the phase curve, including its amplitude and offset. The transmission spectrum rules out clear H$_2$-dominated atmospheres, while secondary atmospheres (made of water, methane, or carbon dioxide) cannot be statistically distinguished from a flat line. The emission spectrum is featureless and consistent with a blackbody-like brightness temperature, helping rule out thick atmospheres with high mean molecular weight. Collecting all evidence, the properties of TOI-1685 b are consistent with a blackbody with no heat redistribution and a low albedo, with a dayside brightness temperature 0.98$\pm$0.07 times that of a perfect blackbody in the NIRSpec NRS2 wavelength range (3.823-5.172 um). Our results add to the growing number of seemingly airless M-star rocky planets, thus constraining the location of the "Cosmic Shoreline". Three independent data reductions have been carried out, all showing a high-amplitude correlated noise component in the white and spectroscopic light curves. The correlated noise properties are different between the NRS1 and NRS2 detectors - importantly the timescales of the strongest components (4.5 hours and 2.5 hours, respectively) - suggesting the noise is from instrumental rather than astrophysical origins. We encourage the community to look into the systematics of NIRSpec for long time-series observations., Comment: 26 pages, 19 figures, 8 tables. Submitted to AAS Journals

Exemplar-based semantic image synthesis aims to generate images aligned with given semantic content while preserving the appearance of an exemplar image. Conventional structure-guidance models, such as ControlNet, are limited in that they cannot directly utilize exemplar images as input, relying instead solely on text prompts to control appearance. Recent tuning-free approaches address this limitation by transferring local appearance from the exemplar image to the synthesized image through implicit cross-image matching in the augmented self-attention mechanism of pre-trained diffusion models. However, these methods face challenges when applied to content-rich scenes with significant geometric deformations, such as driving scenes. In this paper, we propose the Appearance Matching Adapter (AM-Adapter), a learnable framework that enhances cross-image matching within augmented self-attention by incorporating semantic information from segmentation maps. To effectively disentangle generation and matching processes, we adopt a stage-wise training approach. Initially, we train the structure-guidance and generation networks, followed by training the AM-Adapter while keeping the other networks frozen. During inference, we introduce an automated exemplar retrieval method to efficiently select exemplar image-segmentation pairs. Despite utilizing a limited number of learnable parameters, our method achieves state-of-the-art performance, excelling in both semantic alignment preservation and local appearance fidelity. Extensive ablation studies further validate our design choices. Code and pre-trained weights will be publicly available.: https://cvlab-kaist.github.io/AM-Adapter/

We examine the text-free speech representations of raw audio obtained from a self-supervised learning (SSL) model by analyzing the synthesized speech using the SSL representations instead of conventional text representations. Since raw audio does not have paired speech representations as transcribed texts do, obtaining speech representations from unpaired speech is crucial for augmenting available datasets for speech synthesis. Specifically, the proposed speech synthesis is conducted using discrete symbol representations from the SSL model in comparison with text representations, and analytical examinations of the synthesized speech have been carried out. The results empirically show that using text representations is advantageous for preserving semantic information, while using discrete symbol representations is superior for preserving acoustic content, including prosodic and intonational information., Comment: APSIPA ASC 2024

Labeling LiDAR point clouds is notoriously time-and-energy-consuming, which spurs recent unsupervised 3D representation learning methods to alleviate the labeling burden in LiDAR perception via pretrained weights. Almost all existing work focus on a single frame of LiDAR point cloud and neglect the temporal LiDAR sequence, which naturally accounts for object motion (and their semantics). Instead, we propose TREND, namely Temporal REndering with Neural fielD, to learn 3D representation via forecasting the future observation in an unsupervised manner. Unlike existing work that follows conventional contrastive learning or masked auto encoding paradigms, TREND integrates forecasting for 3D pre-training through a Recurrent Embedding scheme to generate 3D embedding across time and a Temporal Neural Field to represent the 3D scene, through which we compute the loss using differentiable rendering. To our best knowledge, TREND is the first work on temporal forecasting for unsupervised 3D representation learning. We evaluate TREND on downstream 3D object detection tasks on popular datasets, including NuScenes, Once and Waymo. Experiment results show that TREND brings up to 90% more improvement as compared to previous SOTA unsupervised 3D pre-training methods and generally improve different downstream models across datasets, demonstrating that indeed temporal forecasting brings improvement for LiDAR perception. Codes and models will be released.

We propose a Multifaceted Resilient Network(MRNet), a novel architecture developed for medical image-to-image translation that outperforms state-of-the-art methods in MRI-to-CT and MRI-to-MRI conversion. MRNet leverages the Segment Anything Model (SAM) to exploit frequency-based features to build a powerful method for advanced medical image transformation. The architecture extracts comprehensive multiscale features from diverse datasets using a powerful SAM image encoder and performs resolution-aware feature fusion that consistently integrates U-Net encoder outputs with SAM-derived features. This fusion optimizes the traditional U-Net skip connection while leveraging transformer-based contextual analysis. The translation is complemented by an innovative dual-mask configuration incorporating dynamic attention patterns and a specialized loss function designed to address regional mapping mismatches, preserving both the gross anatomy and tissue details. Extensive validation studies have shown that MRNet outperforms state-of-the-art architectures, particularly in maintaining anatomical fidelity and minimizing translation artifacts., Comment: This work has been submitted to the IEEE for possible publication

We present a quantum optics-based detection method for determining the position and current of an electron beam. As electrons pass through a dilute vapor of rubidium atoms, their magnetic field perturb the atomic spin's quantum state and causes polarization rotation of a laser resonant with an optical transition of the atoms. By measuring the polarization rotation angle across the laser beam, we recreate a 2D projection of the magnetic field and use it to determine the e-beam position, size and total current. We tested this method for an e-beam with currents ranging from 30 to 110 {\mu}A. Our approach is insensitive to electron kinetic energy, and we confirmed that experimentally between 10 to 20 keV. This technique offers a unique platform for non-invasive characterization of charged particle beams used in accelerators for particle and nuclear physics research.

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

Autonomous agents operating in adversarial scenarios face a fundamental challenge: while they may know their adversaries' high-level objectives, such as reaching specific destinations within time constraints, the exact policies these adversaries will employ remain unknown. Traditional approaches address this challenge by treating the adversary's state as a partially observable element, leading to a formulation as a Partially Observable Markov Decision Process (POMDP). However, the induced belief-space dynamics in a POMDP require knowledge of the system's transition dynamics, which, in this case, depend on the adversary's unknown policy. Our key observation is that while an adversary's exact policy is unknown, their behavior is necessarily constrained by their mission objectives and the physical environment, allowing us to characterize the space of possible behaviors without assuming specific policies. In this paper, we develop Task-Aware Behavior Fields (TAB-Fields), a representation that captures adversary state distributions over time by computing the most unbiased probability distribution consistent with known constraints. We construct TAB-Fields by solving a constrained optimization problem that minimizes additional assumptions about adversary behavior beyond mission and environmental requirements. We integrate TAB-Fields with standard planning algorithms by introducing TAB-conditioned POMCP, an adaptation of Partially Observable Monte Carlo Planning. Through experiments in simulation with underwater robots and hardware implementations with ground robots, we demonstrate that our approach achieves superior performance compared to baselines that either assume specific adversary policies or neglect mission constraints altogether. Evaluation videos and code are available at https://tab-fields.github.io.

We present a general theoretical framework for group resetting dynamics in multi-agent systems in a drift potential. This setup contrasts with a typical resetting problem that involves a single searcher looking for a target, with resetting traditionally studied to optimize the search time to a target. More recently, resetting has also been used as a regulatory mechanism to avoid adverse outcomes, such as preventing critically high water levels in dams or deleveraging financial portfolios. Here, we extend current resetting theories to group dynamics, with applications ranging from bacterial evolution under antibiotic pressure to multiple-searcher optimization algorithms. Our framework incorporates extreme value statistics and renewal theory, from which we derive a master equation for the center of mass distribution of a group of searchers. This master equation allows us to calculate essential observables analytically. For example, how the group's average position depends on group size, resetting rates, drift potential strength, and diffusion constants. This theoretical approach offers a new perspective on optimizing group search and regulatory mechanisms through resetting., Comment: 5 pages, 3 figures

Recent text-to-image diffusion models leverage cross-attention layers, which have been effectively utilized to enhance a range of visual generative tasks. However, our understanding of cross-attention layers remains somewhat limited. In this study, we present a method for constructing Head Relevance Vectors (HRVs) that align with useful visual concepts. An HRV for a given visual concept is a vector with a length equal to the total number of cross-attention heads, where each element represents the importance of the corresponding head for the given visual concept. We develop and employ an ordered weakening analysis to demonstrate the effectiveness of HRVs as interpretable features. To demonstrate the utility of HRVs, we propose concept strengthening and concept adjusting methods and apply them to enhance three visual generative tasks. We show that misinterpretations of polysemous words in image generation can be corrected in most cases, five challenging attributes in image editing can be successfully modified, and catastrophic neglect in multi-concept generation can be mitigated. Overall, our work provides an advancement in understanding cross-attention layers and introduces new approaches for fine-controlling these layers at the head level.

The non-isomorphic locus of a general projection from an embedded smooth projective variety to a hypersurface moves in a linear system of an effective divisor which we call the double point divisor. David Mumford proved that the double point divisor from outer projection is always base point free, and Bo Ilic proved that it is ample except for a Roth variety. The first aim of this paper is to show that the double point divisor from outer projection is very ample except in the Roth case. This answers a question of Bo Ilic. Unlike the case of outer projection, the double point divisor from inner projection may not be base point free nor ample. However, Atsushi Noma proved that it is semiample except when a variety is neither a Roth variety, a scroll over a curve, nor the second Veronese surface. In this paper, we investigate when the double point divisor from inner projection is base point free or big., Comment: 33 pages, comments are welcome

Recent advancements in video large multimodal models (LMMs) have significantly improved their video understanding and reasoning capabilities. However, their performance drops on out-of-distribution (OOD) tasks that are underrepresented in training data. Traditional methods like fine-tuning on OOD datasets are impractical due to high computational costs. While In-context learning (ICL) with demonstration examples has shown promising generalization performance in language tasks and image-language tasks without fine-tuning, applying ICL to video-language tasks faces challenges due to the limited context length in Video LMMs, as videos require longer token lengths. To address these issues, we propose VideoICL, a novel video in-context learning framework for OOD tasks that introduces a similarity-based relevant example selection strategy and a confidence-based iterative inference approach. This allows to select the most relevant examples and rank them based on similarity, to be used for inference. If the generated response has low confidence, our framework selects new examples and performs inference again, iteratively refining the results until a high-confidence response is obtained. This approach improves OOD video understanding performance by extending effective context length without incurring high costs. The experimental results on multiple benchmarks demonstrate significant performance gains, especially in domain-specific scenarios, laying the groundwork for broader video comprehension applications. Code will be released at https://github.com/KangsanKim07/VideoICL

Blind quantum computing (BQC) is a promising application of distributed quantum systems, where a client can perform computations on a remote server without revealing any details of the applied circuit. While the most promising realizations of quantum computers are based on various matter qubit platforms, implementing BQC on matter qubits remains an outstanding challenge. Using silicon-vacancy (SiV) centers in nanophotonic diamond cavities with an efficient optical interface, we experimentally demonstrate a universal quantum gate set consisting of single- and two-qubit blind gates over a distributed two-node network. Using these ingredients, we perform a distributed algorithm with blind operations across our two-node network, paving the way towards blind quantum computation with matter qubits in distributed, modular architectures., Comment: 47 pages, 29 figures

Molecular Relational Learning (MRL) is a rapidly growing field that focuses on understanding the interaction dynamics between molecules, which is crucial for applications ranging from catalyst engineering to drug discovery. Despite recent progress, earlier MRL approaches are limited to using only the 2D topological structure of molecules, as obtaining the 3D interaction geometry remains prohibitively expensive. This paper introduces a novel 3D geometric pre-training strategy for MRL (3DMRL) that incorporates a 3D virtual interaction environment, overcoming the limitations of costly traditional quantum mechanical calculation methods. With the constructed 3D virtual interaction environment, 3DMRL trains 2D MRL model to learn the overall 3D geometric information of molecular interaction through contrastive learning. Moreover, fine-grained interaction between molecules is learned through force prediction loss, which is crucial in understanding the wide range of molecular interaction processes. Extensive experiments on various tasks using real-world datasets, including out-of-distribution and extrapolation scenarios, demonstrate the effectiveness of 3DMRL, showing up to a 24.93\% improvement in performance across 40 tasks.

DEtection TRansformer (DETR) has emerged as a promising architecture for object detection, offering an end-to-end prediction pipeline. In practice, however, DETR generates hundreds of predictions that far outnumber the actual number of objects present in an image. This raises the question: can we trust and use all of these predictions? Addressing this concern, we present empirical evidence highlighting how different predictions within the same image play distinct roles, resulting in varying reliability levels across those predictions. More specifically, while multiple predictions are often made for a single object, our findings show that most often one such prediction is well-calibrated, and the others are poorly calibrated. Based on these insights, we demonstrate identifying a reliable subset of DETR's predictions is crucial for accurately assessing the reliability of the model at both object and image levels. Building on this viewpoint, we first tackle the shortcomings of widely used performance and calibration metrics, such as average precision and various forms of expected calibration error. Specifically, they are inadequate for determining which subset of DETR's predictions should be trusted and utilized. In response, we present Object-level Calibration Error (OCE), which is capable of assessing the calibration quality both across different models and among various configurations within a specific model. As a final contribution, we introduce a post hoc Uncertainty Quantification (UQ) framework that predicts the accuracy of the model on a per-image basis. By contrasting the average confidence scores of positive (i.e., likely to be matched) and negative predictions determined by OCE, the framework assesses the reliability of the DETR model for each test image.

Multimodal Large Language Models (MLLMs) have become increasingly important due to their state-of-the-art performance and ability to integrate multiple data modalities, such as text, images, and audio, to perform complex tasks with high accuracy. This paper presents a comprehensive survey on personalized multimodal large language models, focusing on their architecture, training methods, and applications. We propose an intuitive taxonomy for categorizing the techniques used to personalize MLLMs to individual users, and discuss the techniques accordingly. Furthermore, we discuss how such techniques can be combined or adapted when appropriate, highlighting their advantages and underlying rationale. We also provide a succinct summary of personalization tasks investigated in existing research, along with the evaluation metrics commonly used. Additionally, we summarize the datasets that are useful for benchmarking personalized MLLMs. Finally, we outline critical open challenges. This survey aims to serve as a valuable resource for researchers and practitioners seeking to understand and advance the development of personalized multimodal large language models.

Density matrix exponentiation (DME) is a quantum algorithm that processes multiple copies of a program state $\sigma$ to realize the Hamiltonian evolution $e^{-i \sigma t}$. While serving as a prototypical sample-based quantum algorithm, DME is a powerful tool for various quantum information processing tasks, such as quantum principal component analysis and Hamiltonian simulation. In this work, we present a detailed sample complexity analysis of DME and sample-based Hamiltonian simulation. In particular, we prove that the sample complexity of DME is no larger than $4t^2/\varepsilon$, where $t$ is the desired evolution time and $\varepsilon$ is the desired imprecision level, as quantified by the normalized diamond distance. We also establish a fundamental lower bound on the sample complexity of sample-based Hamiltonian simulation, which matches our DME sample complexity bound up to a constant multiplicative factor, thereby proving that DME is optimal for sample-based Hamiltonian simulation. Finally, we point out that the DME sample complexity analysis in Appendix A of [Kimmel et al., npj Quantum Information 3, 13 (2017)] appears to be incomplete, highlighting the need for the results presented here, given the extensive use of DME over the past decade since its original proposal., Comment: 20 pages

Research on manipulating materials using light has garnered significant interest, yet examples of controlling electronic polarization in magnetic materials remain scarce. Here, we demonstrate the hysteresis of electronic polarization in the antiferromagnetic semiconductor FePS3 via light. Below the N\'eel temperature, we observe linear dichroism (i.e., optical anisotropy) without structural symmetry breaking. Light-induced net polarization aligns along the a-axis (zigzag direction) at 1.6 eV due to the dipolar polarization and along the b-axis (armchair direction) at 2.0 eV due to the combined effects of dipolar and octupolar polarizations, resulting from charge transfer from the armchair to the zigzag direction by light. Unexpected hysteresis of the electronic polarization occurs at 2.0 eV due to the octupolar polarization, in contrast to the absence of such hysteresis at 1.6 eV. We attribute this to a symmetry breaking of the light-induced phase of FePS3 involving electronic polarization within the spin lattice. This study suggests a new mechanism for generating and controlling electronic polarization in magnetic materials using light, with implications for future device applications., Comment: 34 pages, 5 figures, 13 supplementary figures

In many settings, machine learning models may be used to inform decisions that impact individuals or entities who interact with the model. Such entities, or agents, may game model decisions by manipulating their inputs to the model to obtain better outcomes and maximize some utility. We consider a multi-agent setting where the goal is to identify the "worst offenders:" agents that are gaming most aggressively. However, identifying such agents is difficult without knowledge of their utility function. Thus, we introduce a framework in which each agent's tendency to game is parameterized via a scalar. We show that this gaming parameter is only partially identifiable. By recasting the problem as a causal effect estimation problem where different agents represent different "treatments," we prove that a ranking of all agents by their gaming parameters is identifiable. We present empirical results in a synthetic data study validating the usage of causal effect estimation for gaming detection and show in a case study of diagnosis coding behavior in the U.S. that our approach highlights features associated with gaming., Comment: 38 pages, 31 figures. NeurIPS 2024

In certain three-dimensional (3D) applications of photoacoustic computed tomography (PACT), including \textit{in vivo} breast imaging, hemispherical measurement apertures that enclose the object within their convex hull are employed for data acquisition. Data acquired with such measurement geometries are referred to as \textit{half-scan} data, as only half of a complete spherical measurement aperture is employed. Although previous studies have demonstrated that half-scan data can uniquely and stably reconstruct the sought-after object, no closed-form reconstruction formula for use with half-scan data has been reported. To address this, a semi-analytic reconstruction method in the form of filtered backprojection (FBP), referred to as the half-scan FBP method, is developed in this work. Because the explicit form of the filtering operation in the half-scan FBP method is not currently known, a learning-based method is proposed to approximate it. The proposed method is systematically investigated by use of virtual imaging studies of 3D breast PACT that employ ensembles of numerical breast phantoms and a physics-based model of the data acquisition process. The method is subsequently applied to experimental data acquired in an \textit{in vivo} breast PACT study. The results confirm that the half-scan FBP method can accurately reconstruct 3D images from half-scan data. Importantly, because the sought-after inverse mapping is well-posed, the reconstruction method remains accurate even when applied to data that differ considerably from those employed to learn the filtering operation.

As blockchain technology gains traction for enhancing data security and operational efficiency, traditional centralized authentication systems remain a significant bottleneck. This paper addresses the challenge of integrating decentralized authentication and access control within distributed networks. We propose a novel solution named ChainGuard, a fully decentralized authentication and access control mechanism based on smart contracts. ChainGuard eliminates the need for a central server by leveraging blockchain technology to manage user roles and permissions dynamically. Our scheme supports user interactions across multiple organizations simultaneously, enhancing security, efficiency, and transparency. By addressing key challenges such as scalability, security, and transparency, ChainGuard not only bridges the gap between traditional centralized systems and blockchain's decentralized ethos but also enhances data protection and operational efficiency., Comment: accepted at the 2025 IEEE Consumer Communications & Networking Conference (CCNC 2025)

This paper explores null elements in English, Chinese, and Korean Penn treebanks. Null elements contain important syntactic and semantic information, yet they have typically been treated as entities to be removed during language processing tasks, particularly in constituency parsing. Thus, we work towards the removal and, in particular, the restoration of null elements in parse trees. We focus on expanding a rule-based approach utilizing linguistic context information to Chinese, as rule based approaches have historically only been applied to English. We also worked to conduct neural experiments with a language agnostic sequence-to-sequence model to recover null elements for English (PTB), Chinese (CTB) and Korean (KTB). To the best of the authors' knowledge, null elements in three different languages have been explored and compared for the first time. In expanding a rule based approach to Chinese, we achieved an overall F1 score of 80.00, which is comparable to past results in the CTB. In our neural experiments we achieved F1 scores up to 90.94, 85.38 and 88.79 for English, Chinese, and Korean respectively with functional labels., Comment: 10 pages

Pluralistic Image Inpainting (PII) offers multiple plausible solutions for restoring missing parts of images and has been successfully applied to various applications including image editing and object removal. Recently, VQGAN-based methods have been proposed and have shown that they significantly improve the structural integrity in the generated images. Nevertheless, the state-of-the-art VQGAN-based model PUT faces a critical challenge: degradation of detail quality in output images due to feature quantization. Feature quantization restricts the latent space and causes information loss, which negatively affects the detail quality essential for image inpainting. To tackle the problem, we propose the FDM (Feature Dequantization Module) specifically designed to restore the detail quality of images by compensating for the information loss. Furthermore, we develop an efficient training method for FDM which drastically reduces training costs. We empirically demonstrate that our method significantly enhances the detail quality of the generated images with negligible training and inference overheads.

Deep neural network (DNN)-based policy models like vision-language-action (VLA) models are transformative in automating complex decision-making across applications by interpreting multi-modal data. However, scaling these models greatly increases computational costs, which presents challenges in fields like robot manipulation and autonomous driving that require quick, accurate responses. To address the need for deployment on resource-limited hardware, we propose a new quantization framework for IL-based policy models that fine-tunes parameters to enhance robustness against low-bit precision errors during training, thereby maintaining efficiency and reliability under constrained conditions. Our evaluations with representative robot manipulation for 4-bit weight-quantization on a real edge GPU demonstrate that our framework achieves up to 2.5x speedup and 2.5x energy savings while preserving accuracy. For 4-bit weight and activation quantized self-driving models, the framework achieves up to 3.7x speedup and 3.1x energy saving on a low-end GPU. These results highlight the practical potential of deploying IL-based policy models on resource-constrained devices.