Search

Vision transformers dominate image processing tasks due to their superior performance. However, the quadratic complexity of self-attention limits the scalability of these systems and their deployment on resource-constrained devices. State Space Models (SSMs) have emerged as a solution by introducing a linear recurrence mechanism, which reduces the complexity of sequence modeling from quadratic to linear. Recently, SSMs have been extended to high-resolution vision tasks. Nonetheless, the linear recurrence mechanism struggles to fully utilize matrix multiplication units on modern hardware, resulting in a computational bottleneck. We address this issue by introducing \textit{VMeanba}, a training-free compression method that eliminates the channel dimension in SSMs using mean operations. Our key observation is that the output activations of SSM blocks exhibit low variances across channels. Our \textit{VMeanba} leverages this property to optimize computation by averaging activation maps across the channel to reduce the computational overhead without compromising accuracy. Evaluations on image classification and semantic segmentation tasks demonstrate that \textit{VMeanba} achieves up to a 1.12x speedup with less than a 3\% accuracy loss. When combined with 40\% unstructured pruning, the accuracy drop remains under 3\%., Comment: Accepted by NeurIPS 2024 Machine Learning for Systems workshop

Multimodal foundation models, such as Gemini and ChatGPT, have revolutionized human-machine interactions by seamlessly integrating various forms of data. Developing a universal spoken language model that comprehends a wide range of natural language instructions is critical for bridging communication gaps and facilitating more intuitive interactions. However, the absence of a comprehensive evaluation benchmark poses a significant challenge. We present Dynamic-SUPERB Phase-2, an open and evolving benchmark for the comprehensive evaluation of instruction-based universal speech models. Building upon the first generation, this second version incorporates 125 new tasks contributed collaboratively by the global research community, expanding the benchmark to a total of 180 tasks, making it the largest benchmark for speech and audio evaluation. While the first generation of Dynamic-SUPERB was limited to classification tasks, Dynamic-SUPERB Phase-2 broadens its evaluation capabilities by introducing a wide array of novel and diverse tasks, including regression and sequence generation, across speech, music, and environmental audio. Evaluation results indicate that none of the models performed well universally. SALMONN-13B excelled in English ASR, while WavLLM demonstrated high accuracy in emotion recognition, but current models still require further innovations to handle a broader range of tasks. We will soon open-source all task data and the evaluation pipeline.

Automated grading has become an essential tool in education technology due to its ability to efficiently assess large volumes of student work, provide consistent and unbiased evaluations, and deliver immediate feedback to enhance learning. However, current systems face significant limitations, including the need for large datasets in few-shot learning methods, a lack of personalized and actionable feedback, and an overemphasis on benchmark performance rather than student experience. To address these challenges, we propose a Zero-Shot Large Language Model (LLM)-Based Automated Assignment Grading (AAG) system. This framework leverages prompt engineering to evaluate both computational and explanatory student responses without requiring additional training or fine-tuning. The AAG system delivers tailored feedback that highlights individual strengths and areas for improvement, thereby enhancing student learning outcomes. Our study demonstrates the system's effectiveness through comprehensive evaluations, including survey responses from higher education students that indicate significant improvements in motivation, understanding, and preparedness compared to traditional grading methods. The results validate the AAG system's potential to transform educational assessment by prioritizing learning experiences and providing scalable, high-quality feedback.

Emerging two-dimensional transition metal dichalcogenides (TMDCs) offer a promising platform for on-chip integrated photonics because of their unique optical and electronic properties. Their naturally passivated surfaces make them highly tolerant to lattice mismatch, enabling seamless heterogeneous integration by stacking different van der Waals materials, a crucial step in the development of advanced photonic devices. Here, we demonstrate the use of an inverse design wavelength demultiplexing waveguides for on-chip sorting and routing of distinct photoluminescence from the heterojunction formed by WS2 and WSe2 monolayers. The integrated nanophotonic chip splits and sorts excitonic emission into individual waveguides at room temperature. Our demonstration opens up new perspectives for integrating light sources in van der Waals materials with functional integrated photonics, offering a versatile platform for both fundamental research and practical applications.

The rapid proliferation of large language models (LLMs) in natural language processing (NLP) has created a critical need for techniques that enable efficient deployment on memory-constrained devices without compromising performance. We present a method to prune LLMs that selectively prunes model blocks based on an importance score and replaces them with a low-parameter replacement strategy. Specifically, we propose a principled metric to replace each pruned block using a weight-sharing mechanism that leverages unpruned counterparts from the model and block-specific low-rank adapters. Furthermore, we facilitate the learning of these replacement blocks with output feature normalization and an adapter initialization scheme built on low-rank SVD reconstructions. Empirical evaluations demonstrate substantial performance gains over existing methods, achieving state-of-the-art performance on 5/6 benchmarks for a compression rate of 30% and 6/6 benchmarks for a compression rate of 40%. We also demonstrate that our approach can extend smaller models, boosting performance on 6/6 benchmarks using only ~0.3% tokens of extended training with minimal additional parameter costs., Comment: Accepted to NAACL 2025 - Main Conference

We develop Monte Carlo simulations for uniformly charged polymers and machine learning algorithm to interpret the intra-polymer structure factor of the charged polymer system, which can be obtained from small-angle scattering experiments. The polymer is modeled as a chain of fixed-length bonds, where the connected bonds are subject to bending energy, and there is also a screened Coulomb potential for charge interaction between all joints. The bending energy is determined by the intrinsic bending stiffness, and the charge interaction depends on the interaction strength and screening length. All three contribute to the stiffness of the polymer chain and lead to longer and larger polymer conformations. The screening length also introduces a second length scale for the polymer besides the bending persistence length. To obtain the inverse mapping from the structure factor to these polymer conformation and energy-related parameters, we generate a large data set of structure factors by running simulations for a wide range of polymer energy parameters. We use principal component analysis to investigate the intra-polymer structure factors and determine the feasibility of the inversion using the nearest neighbor distance. We employ Gaussian process regression to achieve the inverse mapping and extract the characteristic parameters of polymers from the structure factor with low relative error., Comment: 8 pages, 8 figures

For a positive integer $m$, a finite group $G$ is said to admit a tournament $m$-semiregular representation (TmSR for short) if there exists a tournament $\Gamma$ such that the automorphism group of $\Gamma$ is isomorphic to $G$ and acts semiregularly on the vertex set of $\Gamma$ with $m$ orbits. Clearly, every finite group of even order does not admit a TmSR for any positive integer $m$, and T1SR is the well-known tournament regular representation (TRR for short). In 1986, Godsil \cite{god} proved, by a probabilistic approach, that the only finite groups of odd order without a TRR are $\mathbb{Z}_3^2$ and $\mathbb{Z}_3^3$ . More recently, Du \cite{du} proved that every finite group of odd order has a TmSR for every $m \geq 2$. The author of \cite{du} observed that a finite group of odd order has no regular TmSR when $m$ is an even integer, a group of order $1$ has no regular T3SR, and $\mathbb{Z}_3^2$ admits a regular T3SR. At the end of \cite{du}, Du proposed the following problem. \noindent{\sf\it Problem.} \ \ {\it For every odd integer $m\geq 3$, classify finite groups of odd order which have a regular TmSR.} The motivation of this paper is to give an answer for the above problem. We proved that if $G$ is a finite group with odd order $n>1$, then $G$ admits a regular TmSR for any odd integer $m\geq 3$.

We investigate the longstanding problem of determining the maximum size of a $(d+1)$-uniform set system with VC-dimension at most $d$. Since the seminal 1984 work of Frankl and Pach, which established the elegant upper bound $\binom{n}{d}$, this question has resisted significant progress. The best-known lower bound is $\binom{n-1}{d} + \binom{n-4}{d-2}$, obtained by Ahlswede and Khachatrian, leaving a substantial gap of $\binom{n-1}{d-1}-\binom{n-4}{d-2}$. Despite decades of effort, improvements to the Frankl--Pach bound have been incremental at best: Mubayi and Zhao introduced an $\Omega_d(\log{n})$ improvement for prime powers $d$, while Ge, Xu, Yip, Zhang, and Zhao achieved a gain of 1 for general $d$. In this work, we provide a purely combinatorial approach that significantly sharpens the Frankl--Pach upper bound. Specifically, for large $n$, we demonstrate that the Frankl--Pach bound can be improved to $\binom{n}{d} - \binom{n-1}{d-1} + O_d(n^{d-1 - \frac{1}{4d-2}})=\binom{n-1}{d}+O_d(n^{d-1 - \frac{1}{4d-2}})$. This result completely removes the main term $\binom{n-1}{d-1}$ from the previous gap between the known lower and upper bounds. It also offers fresh insights into the combinatorial structure of uniform set systems with small VC-dimension. In addition, the original Erd\H{o}s--Frankl--Pach conjecture, which sought to generalize the EKR theorem in the 1980s, has been disproven. We propose a new refined conjecture that might establish a sturdier bridge between VC-dimension and the EKR theorem, and we verify several specific cases of this conjecture, which is of independent interest., Comment: 25 pages

We present compelling evidence that Dark Matter (DM)-neutrino interactions can resolve the persistent structure growth parameter discrepancy, $S_8 = \sigma_8\,\sqrt{\Omega_m/0.3}$, between early and late universe observations. By incorporating cosmic shear measurements from current Weak Lensing (WL) surveys and leveraging an emulator based on true $N$-body simulations to account for nonlinear corrections, we demonstrate that an interaction strength of $u \sim 10^{-4}$ not only provides a coherent explanation for the high-multipole observations from the Atacama Cosmology Telescope (ACT), but also alleviates the $S_8$ discrepancy. Combining early universe constraints with DES Y3 cosmic shear data yields a nearly $3\sigma$ detection of non-zero DM-neutrino interactions, marking a significant breakthrough in cosmological research. Our findings challenge the standard $\Lambda$CDM paradigm and highlight the potential of future large-scale structure surveys, which can rigorously test this interaction and unveil the fundamental properties of dark matter., Comment: 28 pages, 10 figures

Hypothesis Understanding the scission of rod-like micelles under mechanical forces is crucial for optimizing their stability and behavior in industrial applications. This study investigates how micelle length, flexibility, and external forces interact, offering insights into the design of micellar systems in processes influenced by mechanical stress. Although significant, direct experimental observations of flow-induced micellar scission using scattering techniques remain scarce. Experiments and Simulations Small angle neutron scattering (SANS) is used to explore the shear response of aqueous cetyltrimethylammonium bromide (CTAB) solutions with sodium nitrate. Rheological tests show shear thinning with no shear banding, ensuring a uniform flow field for reliable interpretation of scattering data. As shear rate increases, the scattering spectra show angular distortion, which is analyzed using spherical harmonic decomposition to characterize flow-induced scission and micelle orientation under shear. Findings Two analysis steps are used: a model-independent spectral eigendecomposition reveals a decrease in micellar length, while regression analysis quantifies the evolution of the length distribution and mean length with shear rate. Additionally, micelle alignment increases with shear, quantified by the orientational distribution function. These findings provide experimental evidence for flow-induced alignment and scission, offering a new framework for understanding shear-induced phenomena in micellar systems

Few-shot Semantic Segmentation (FSS) is a challenging task that utilizes limited support images to segment associated unseen objects in query images. However, recent FSS methods are observed to perform worse, when enlarging the number of shots. As the support set enlarges, existing FSS networks struggle to concentrate on the high-contributed supports and could easily be overwhelmed by the low-contributed supports that could severely impair the mask predictions. In this work, we study this challenging issue, called support dilution, our goal is to recognize, select, preserve, and enhance those high-contributed supports in the raw support pool. Technically, our method contains three novel parts. First, we propose a contribution index, to quantitatively estimate if a high-contributed support dilutes. Second, we develop the Symmetric Correlation (SC) module to preserve and enhance the high-contributed support features, minimizing the distraction by the low-contributed features. Third, we design the Support Image Pruning operation, to retrieve a compact and high quality subset by discarding low-contributed supports. We conduct extensive experiments on two FSS benchmarks, COCO-20i and PASCAL-5i, the segmentation results demonstrate the compelling performance of our solution over state-of-the-art FSS approaches. Besides, we apply our solution for online segmentation and real-world segmentation, convincing segmentation results showing the practical ability of our work for real-world demonstrations., Comment: 15 pages, 15 figures

We prove Gamma conjecture I for any flag varieties by following a strategy proposed by Galkin and Iritani. The main ingredient is to prove that the $\widehat{\Gamma}$-class of a flag variety is mirror to the totally positive part of the corresponding Rietsch mirror.

Weexperimentally realize the continuously controllable dissipative coupling and coherent coupling induced by different magnon modes and the same anti-resonance. It has been observed that the weaker the microwave magnetic field distribution of the magnon mode in magnetic materials, the more likely dissipative coupling is to occur. Conversely, stronger magnetic field distributions favor coherent coupling. Based on this principle, we have designed and implemented a system that alternates between dissipative and coherent coupling regimes. It allows microwave signals to be selectively transmitted over a large applied magnetic field range at the frequency of anti-resonance. Our experimental achievements may promote the construction of new magnonics devices like magnetic-tuning switch., Comment: 7 pages, 6 figures

Molecular orbital (MO) is one of the most fundamental concepts for molecules, relating to all branches of chemistry, while scanning tunneling microscopy (STM) has been widely recognized for its potential to measure the spatial distribution of MOs. However, the precise characterization of MO with high resolution in real space is a long-standing challenge owing to the inevitable interference of high-angular-momentum contributions from functionalized tips in STM. Here, leveraging advances in artificial intelligence for image recognition, we establish a physics-driven deep-learning network, named STM-Net, to reconstruct MOs from high-resolution STM images with a functionalized tip, taking advantage of the separable characteristics of different angular momentum contributions. We demonstrate that STM-Net can be directly applied to a variety of experimental observations, successfully reconstructing pristine MO features for molecules under diverse conditions. Moreover, STM-Net can adapt to various states of the functionalized tip and the substrate, illustrating the broad applicability of our physics-driven framework. These results pave the way for accurate characterization of MO with high resolution, potentially leading to new insights and applications for this fundamental concept in chemistry., Comment: 19 pages, 4 figures, 4 extended data figures

Survival prognosis is crucial for medical informatics. Practitioners often confront small-sized clinical data, especially cancer patient cases, which can be insufficient to induce useful patterns for survival predictions. This study deals with small sample survival analysis by leveraging transfer learning, a useful machine learning technique that can enhance the target analysis with related knowledge pre-learned from other data. We propose and develop various transfer learning methods designed for common survival models. For parametric models such as DeepSurv, Cox-CC (Cox-based neural networks), and DeepHit (end-to-end deep learning model), we apply standard transfer learning techniques like pretraining and fine-tuning. For non-parametric models such as Random Survival Forest, we propose a new transfer survival forest (TSF) model that transfers tree structures from source tasks and fine-tunes them with target data. We evaluated the transfer learning methods on colorectal cancer (CRC) prognosis. The source data are 27,379 SEER CRC stage I patients, and the target data are 728 CRC stage I patients from the West China Hospital. When enhanced by transfer learning, Cox-CC's $C^{td}$ value was boosted from 0.7868 to 0.8111, DeepHit's from 0.8085 to 0.8135, DeepSurv's from 0.7722 to 0.8043, and RSF's from 0.7940 to 0.8297 (the highest performance). All models trained with data as small as 50 demonstrated even more significant improvement. Conclusions: Therefore, the current survival models used for cancer prognosis can be enhanced and improved by properly designed transfer learning techniques. The source code used in this study is available at https://github.com/YonghaoZhao722/TSF.

We study the Scholze test functions for bad reduction of simple Shimura varieties at a prime where the underlying local group is any inner form of a product of Weil restrictions of general linear groups. Using global methods, we prove that these test functions satisfy a vanishing property of their twisted orbital integrals, and we prove that the pseudostabilization base changes of such functions exist (even though the local group need not be quasi-split) and can be expressed in terms of explicit distributions in the stable Bernstein center. We then deduce applications to the stable trace formula and local Hasse-Weil zeta functions for these Shimura varieties.

The rapid rise of real-time communication and large language models has significantly increased the importance of speech compression. Deep learning-based neural speech codecs have outperformed traditional signal-level speech codecs in terms of rate-distortion (RD) performance. Typically, these neural codecs employ an encoder-quantizer-decoder architecture, where audio is first converted into latent code feature representations and then into discrete tokens. However, this architecture exhibits insufficient RD performance due to two main drawbacks: (1) the inadequate performance of the quantizer, challenging training processes, and issues such as codebook collapse; (2) the limited representational capacity of the encoder and decoder, making it difficult to meet feature representation requirements across various bitrates. In this paper, we propose a rate-aware learned speech compression scheme that replaces the quantizer with an advanced channel-wise entropy model to improve RD performance, simplify training, and avoid codebook collapse. We employ multi-scale convolution and linear attention mixture blocks to enhance the representational capacity and flexibility of the encoder and decoder. Experimental results demonstrate that the proposed method achieves state-of-the-art RD performance, obtaining 53.51% BD-Rate bitrate saving in average, and achieves 0.26 BD-VisQol and 0.44 BD-PESQ gains.

Muon-spin-rotation ($\mu$SR) experiments and the observation of a spontaneous Nernst effect indicate time-reversal symmetry breaking (BTRS) at $T_{\rm c}^{\rm Z2}$ above the superconducting transition temperature $T_{\rm c}$ in Ba$_{1-x}$K$_x$Fe$_2$As$_2$, with $x\approx0.8$. Further studies have pointed out that BTRS is caused by the formation of a new state of matter associated with the condensation of pairs of electron pairs. Despite exhibiting multiple unconventional effects that warrant further investigation, the electronic spectral properties of this electron quadrupling state remain largely unexplored. Here, we present detailed $^{75}$As nuclear magnetic resonance (NMR) measurements of Ba$_{1-x}$K$_x$Fe$_2$As$_2$, with $x = 0.77$, which has $T_{\rm c}^{\rm Z2}$ > $T_{\rm c}$ according to measurements of the spontaneous Nernst effect. The NMR data obtained in this work provide the first direct electronic spectral characteristics of the electron quadrupling state by indicating that it evolves from a pseudogap that sets in at $T^*$ well above $T_{\rm c}^{\rm Z2}$. This pseudogap behavior is consistent with $\mu$SR Knight-shift, specific-heat, and transport data indicating the formation of a bound state of electrons. According to a theory of electron quadrupling condensates, such bound-state formations should precede the onset of BTRS correlations between pairs of electron pairs. The second important insight from NMR data is the absence of spin-related magnetism. The temperature dependence of the spin-lattice relaxation rate $1/T_1T$ and the evolution of the NMR linewidth prove the absence of a magnetic transition at $T_{\rm c}^{\rm Z2}$ and rule out even a proximity to some magnetic instability. This indicates that the spontaneous magnetic fields detected in this compound are not caused by spin magnetism but are associated with persistent real-space currents., Comment: 22 pages, 18 figures

We compare three filtrations of the tope space of an oriented matroid. The first is the dual Varchenko-Gelfand degree filtration, the second filtration is from Kalinin's spectral sequence, and the last one derives from Quillen's augmentation filtration. We show that all three filtrations and the respective maps coincide over $\mathbb{Z}/ 2\mathbb{Z}$. We also show that the dual Varchenko-Gelfand degree filtration can be made into a filtration of the $\mathbb{Z}$-sign cosheaf on the fan of the underlying matroid. This was previously carried out with $\mathbb{Z}/ 2\mathbb{Z}$-coefficients by the first author and Renaudineau using the Quillen filtration and has applications to real algebraic geometry via patchworking.

Empirically, Deep Learning (DL) has demonstrated unprecedented success in practical applications. However, DL remains by and large a mysterious "black-box", spurring recent theoretical research to build its mathematical foundations. In this paper, we investigate DL for data classification through the prism of metric topology. Considering that conventional Euclidean metric over the network parameter space typically fails to discriminate DL networks according to their classification outcomes, we propose from a probabilistic point of view a meaningful distance measure, whereby DL networks yielding similar classification performances are close. The proposed distance measure defines such an equivalent relation among network parameter vectors that networks performing equally well belong to the same equivalent class. Interestingly, our proposed distance measure can provably serve as a metric on the quotient set modulo the equivalent relation. Then, under quite mild conditions it is shown that, apart from a vanishingly small subset of networks likely to predict non-unique labels, our proposed metric space is compact, and coincides with the well-known quotient topological space. Our study contributes to fundamental understanding of DL, and opens up new ways of studying DL using fruitful metric space theory.

Logic bugs are bugs that can cause database management systems (DBMSs) to silently produce incorrect results for given queries. Such bugs are severe, because they can easily be overlooked by both developers and users, and can cause applications that rely on the DBMSs to malfunction. In this work, we propose Constant-Optimization-Driven Database Testing (CODDTest) as a novel approach for detecting logic bugs in DBMSs. This method draws inspiration from two well-known optimizations in compilers: constant folding and constant propagation. Our key insight is that for a certain database state and query containing a predicate, we can apply constant folding on the predicate by replacing an expression in the predicate with a constant, anticipating that the results of this predicate remain unchanged; any discrepancy indicates a bug in the DBMS. We evaluated CODDTest on five mature and extensively-tested DBMSs-SQLite, MySQL, CockroachDB, DuckDB, and TiDB-and found 45 unique, previously unknown bugs in them. Out of these, 24 are unique logic bugs. Our manual analysis of the state-of-the-art approaches indicates that 11 logic bugs are detectable only by CODDTest. We believe that CODDTest is easy to implement, and can be widely adopted in practice.

Clinical trials are a critical process in the medical field for introducing new treatments and innovations. However, cohort selection for clinical trials is a time-consuming process that often requires manual review of patient text records for specific keywords. Though there have been studies on standardizing the information across the various platforms, Natural Language Processing (NLP) tools remain crucial for spotting eligibility criteria in textual reports. Recently, pre-trained large language models (LLMs) have gained popularity for various NLP tasks due to their ability to acquire a nuanced understanding of text. In this paper, we study the performance of large language models on clinical trial cohort selection and leverage the n2c2 challenges to benchmark their performance. Our results are promising with regard to the incorporation of LLMs for simple cohort selection tasks, but also highlight the difficulties encountered by these models as soon as fine-grained knowledge and reasoning are required.

Reconstructing MR images using deep neural networks from undersampled k-space data without using fully sampled training references offers significant value in practice, which is a self-supervised regression problem calling for effective prior knowledge and supervision. The Siamese architectures are motivated by the definition "invariance" and shows promising results in unsupervised visual representative learning. Building homologous transformed images and avoiding trivial solutions are two major challenges in Siamese-based self-supervised model. In this work, we explore Siamese architecture for MRI reconstruction in a self-supervised training fashion called SiamRecon. We show the proposed approach mimics an expectation maximization algorithm. The alternative optimization provide effective supervision signal and avoid collapse. The proposed SiamRecon achieves the state-of-the-art reconstruction accuracy in the field of self-supervised learning on both single-coil brain MRI and multi-coil knee MRI.

The concepts of linkage, building blocks, and problem decomposition have long existed in the genetic algorithm (GA) field and have guided the development of model-based GAs for decades. However, their definitions are usually vague, making it difficult to develop theoretical support. This paper provides an algorithm-independent definition to describe the concept of linkage. With this definition, the paper proves that any problems with a bounded degree of linkage are decomposable and that proper problem decomposition is possible via linkage learning. The way of decomposition given in this paper also offers a new perspective on nearly decomposable problems with bounded difficulty and building blocks from the theoretical aspect. Finally, this paper relates problem decomposition to PAC learning and proves that the global optima of these problems and the minimum decomposition blocks are PAC learnable under certain conditions.

Various benchmarks have been proposed to assess the performance of large language models (LLMs) in different coding scenarios. We refer to them as code-related benchmarks. However, there are no systematic guidelines by which such a benchmark should be developed to ensure its quality, reliability, and reproducibility. We propose How2Bench, which is comprised of a 55- 55-criteria checklist as a set of guidelines to govern the development of code-related benchmarks comprehensively. Using HOW2BENCH, we profiled 274 benchmarks released within the past decade and found concerning issues. Nearly 70% of the benchmarks did not take measures for data quality assurance; over 10% did not even open source or only partially open source. Many highly cited benchmarks have loopholes, including duplicated samples, incorrect reference codes/tests/prompts, and unremoved sensitive/confidential information. Finally, we conducted a human study involving 49 participants, which revealed significant gaps in awareness of the importance of data quality, reproducibility, and transparency., Comment: 42 pages

Spatial reasoning is an essential problem in embodied AI research. Efforts to enhance spatial reasoning abilities through supplementary spatial data and fine-tuning have proven limited and ineffective when addressing complex embodied tasks, largely due to their dependence on language-based outputs. While some approaches have introduced a point-based action space to mitigate this issue, they fall short in managing more intricate tasks within complex environments. This deficiency arises from their failure to fully exploit the inherent thinking and reasoning capabilities that are fundamental strengths of Vision-Language Models (VLMs). To address these limitations, we propose a novel approach named SpatialCoT, specifically designed to bolster the spatial reasoning capabilities of VLMs. Our approach comprises two stages: spatial coordinate bi-directional alignment, which aligns vision-language inputs with spatial coordinates, and chain-of-thought spatial grounding, which harnesses the reasoning capabilities of language models for advanced spatial reasoning. We evaluate SpatialCoT on challenging navigation and manipulation tasks, both in simulation and real-world settings. Experimental results demonstrate that our method significantly outperforms previous state-of-the-art approaches in both tasks., Comment: Under Review

We present DeepSSM, an open-source code powered by neural networks (NNs) to emulate gravitational wave (GW) spectra produced by sound waves during cosmological first-order phase transitions in the radiation-dominated era. The training data is obtained from an enhanced version of the Sound Shell Model (SSM), which accounts for the effects of cosmic expansion and yields more accurate spectra in the infrared regime. The emulator enables instantaneous predictions of GW spectra given the phase transition parameters, while achieving agreement with the enhanced SSM model within 10\% accuracy in the worst-case scenarios. The emulator is highly computationally efficient and fully differentiable, making it particularly suitable for direct Bayesian inference on phase transition parameters without relying on empirical templates, such as broken power-law models. We demonstrate this capability by successfully reconstructing phase transition parameters and their degeneracies from mock LISA observations using a Hamiltonian Monte Carlo sampler. The code is available at: https://github.com/ctian282/DeepSSM., Comment: 20 pages, 5 figures, 2 tables

We present GeoManip, a framework to enable generalist robots to leverage essential conditions derived from object and part relationships, as geometric constraints, for robot manipulation. For example, cutting the carrot requires adhering to a geometric constraint: the blade of the knife should be perpendicular to the carrot's direction. By interpreting these constraints through symbolic language representations and translating them into low-level actions, GeoManip bridges the gap between natural language and robotic execution, enabling greater generalizability across diverse even unseen tasks, objects, and scenarios. Unlike vision-language-action models that require extensive training, operates training-free by utilizing large foundational models: a constraint generation module that predicts stage-specific geometric constraints and a geometry parser that identifies object parts involved in these constraints. A solver then optimizes trajectories to satisfy inferred constraints from task descriptions and the scene. Furthermore, GeoManip learns in-context and provides five appealing human-robot interaction features: on-the-fly policy adaptation, learning from human demonstrations, learning from failure cases, long-horizon action planning, and efficient data collection for imitation learning. Extensive evaluations on both simulations and real-world scenarios demonstrate GeoManip's state-of-the-art performance, with superior out-of-distribution generalization while avoiding costly model training., Comment: 32 pages, 13 figures

The proliferation of Vision-Language Models (VLMs) in the past several years calls for rigorous and comprehensive evaluation methods and benchmarks. This work analyzes existing VLM evaluation techniques, including automated metrics, AI-based assessments, and human evaluations across diverse tasks. We first introduce Robin - a novel suite of VLMs that we built by combining Large Language Models (LLMs) and Vision Encoders (VEs) at multiple scales, and use Robin to identify shortcomings of current evaluation approaches across scales. Next, to overcome the identified limitations, we introduce CHIRP - a new long form response benchmark we developed for more robust and complete VLM evaluation. We provide open access to the Robin training code, model suite, and CHIRP benchmark to promote reproducibility and advance VLM research.

Large Reconstruction Models (LRMs) have recently become a popular method for creating 3D foundational models. Training 3D reconstruction models with 2D visual data traditionally requires prior knowledge of camera poses for the training samples, a process that is both time-consuming and prone to errors. Consequently, 3D reconstruction training has been confined to either synthetic 3D datasets or small-scale datasets with annotated poses. In this study, we investigate the feasibility of 3D reconstruction using unposed video data of various objects. We introduce UVRM, a novel 3D reconstruction model capable of being trained and evaluated on monocular videos without requiring any information about the pose. UVRM uses a transformer network to implicitly aggregate video frames into a pose-invariant latent feature space, which is then decoded into a tri-plane 3D representation. To obviate the need for ground-truth pose annotations during training, UVRM employs a combination of the score distillation sampling (SDS) method and an analysis-by-synthesis approach, progressively synthesizing pseudo novel-views using a pre-trained diffusion model. We qualitatively and quantitatively evaluate UVRM's performance on the G-Objaverse and CO3D datasets without relying on pose information. Extensive experiments show that UVRM is capable of effectively and efficiently reconstructing a wide range of 3D objects from unposed videos.

Coflow represents a network abstraction that models communication patterns within data centers. The scheduling of coflows is a prevalent issue in large data center environments and is classified as an $\mathcal{NP}$-hard problem. This paper focuses on the scheduling of coflows in heterogeneous parallel networks, defined by architectures featuring multiple network cores operating concurrently. We introduce two pseudo-polynomial-time algorithms and two polynomial-time approximation algorithms to minimize the maximum completion time (makespan) in heterogeneous parallel networks. We propose a randomized algorithm that offers an expected approximation ratio of 1.5. Building upon this foundation, we provide a deterministic algorithm that utilizes derandomization techniques, which offers a performance guarantee of $1.5 + \frac{1}{2 \cdot LB}$, where $LB$ is the lower bound of the makespan for each instance. To address time complexity concerns, we implement an exponential partitioning of time intervals and present a randomized algorithm with an expected approximation ratio of $1.5 + \epsilon$ in polynomial time where $\epsilon>0$. Additionally, we develop a deterministic algorithm with a performance guarantee expressed as $\max\left\{1.5+\epsilon, 1.5+\frac{1}{2 \cdot LB}\right\}$ within polynomial time. These advancements markedly enhance the best-known approximation ratio of $2+\epsilon$.

Atomistic structures of materials offer valuable insights into their functionality. Determining these structures remains a fundamental challenge in materials science, especially for systems with defects. While both experimental and computational methods exist, each has limitations in resolving nanoscale structures. Core-level spectroscopies, such as x-ray absorption (XAS) or electron energy-loss spectroscopies (EELS), have been used to determine the local bonding environment and structure of materials. Recently, machine learning (ML) methods have been applied to extract structural and bonding information from XAS/EELS, but most of these frameworks rely on a single data stream, which is often insufficient. In this work, we address this challenge by integrating multimodal ab initio simulations, experimental data acquisition, and ML techniques for structure characterization. Our goal is to determine local structures and properties using EELS and XAS data from multiple elements and edges. To showcase our approach, we use various lithium nickel manganese cobalt (NMC) oxide compounds which are used for lithium ion batteries, including those with oxygen vacancies and antisite defects, as the sample material system. We successfully inferred local element content, ranging from lithium to transition metals, with quantitative agreement with experimental data. Beyond improving prediction accuracy, we find that ML model based on multimodal spectroscopic data is able to determine whether local defects such as oxygen vacancy and antisites are present, a task which is impossible for single mode spectra or other experimental techniques. Furthermore, our framework is able to provide physical interpretability, bridging spectroscopy with the local atomic and electronic structures.

Chiral perovskites are emerging semiconducting materials with broken symmetry that can selectively absorb and emit circularly polarized light. However, most of the chiral perovskites are typically low-dimensional structures with limited electrical conductivity and their light absorption occurs in the UV region. In this work, we find doping 2,3,5,6-Tetrafluoro-7,7,8,8-tetracyanoquinodimethane (F4TCNQ) in the chiral perovskite matrix can improve the electrical conductivity with an addition of visible light absorption through the emerging charge-transfer electronic states. The new absorption feature exhibits strong circular dichroism adapted from the chiral matrix, which is indicative of a chirality transfer from the host to the guest via an electronic coupling. The charge transfer state is validated by transient absorption spectroscopy and theory modeling. Quantum-chemical modeling identifies a strong wave function overlap between an electron and a hole of the guest-host in a closely packed crystal configuration forming the charge-transfer absorption state. We then integrate the doped chiral perovskite film in photodetectors and demonstrate a selective detection of circularly polarized light both in the UV and visible range. Our results suggest a universal approach of introducing visible photo absorption states to the chiral matrix to broaden the optical active range and enhance the conductivity.

Maintaining a comfortable temperature inside a building requires appropriate thermal insulation of windows, which can be optimised iteratively with numerical simulation. Smoothed particle hydrodynamics(SPH) is a fully Lagrangian method widely used for simulating multi-physics applications with high computational efficiency and accuracy. It is advantageous in physically coupled problems such as heat-fluid-solid or any other type of physically coupled simulations. The focus of this study is to simulate the heat transfer process in various window frames under convective boundary conditions according to ISO10077-2:2012. This paper demonstrates the accuracy and compatibility of SPH when dealing with heat transfer problems, which ensures further development of thermal coupling with other physical fields. The results and methods used in this paper provide some guidance on how to properly handle heat transfer simulations using SPH, which can be extended to multi-physics coupled simulations in the future., Comment: 22 pages, 5 figures and 5 tables

Configuration translation is a critical and frequent task in network operations. When a network device is damaged or outdated, administrators need to replace it to maintain service continuity. The replacement devices may originate from different vendors, necessitating configuration translation to ensure seamless network operation. However, translating configurations manually is a labor-intensive and error-prone process. In this paper, we propose an intent-based framework for translating network configuration with Large Language Model (LLM) Agents. The core of our approach is an Intent-based Retrieval Augmented Generation (IRAG) module that systematically splits a configuration file into fragments, extracts intents, and generates accurate translations. We also design a two-stage verification method to validate the syntax and semantics correctness of the translated configurations. We implement and evaluate the proposed method on real-world network configurations. Experimental results show that our method achieves 97.74% syntax correctness, outperforming state-of-the-art methods in translation accuracy.

Many active galaxies harbor powerful relativistic jets, however, the detailed mechanisms of their formation and acceleration remain poorly understood. To investigate the area of jet acceleration and collimation with the highest available angular resolution, we study the innermost region of the bipolar jet in the nearby low-ionization nuclear emission-line region (LINER) galaxy NGC 1052. We combined observations of NGC 1052 taken with VLBA, GMVA, and EHT over one week in the spring of 2017. For the first time, NGC 1052 was detected with the EHT, providing a size of the central region in-between both jet bases of 250 RS (Schwarzschild radii) perpendicular to the jet axes. This size estimate supports previous studies of the jets expansion profile which suggest two breaks of the profile at around 300 RS and 10000 RS distances to the core. Furthermore, we estimated the magnetic field to be 1.25 Gauss at a distance of 22 {\mu}as from the central engine by fitting a synchrotron-self absorption spectrum to the innermost emission feature, which shows a spectral turn-over at about 130 GHz. Assuming a purely poloidal magnetic field, this implies an upper limit on the magnetic field strength at the event horizon of 26000 Gauss, which is consistent with previous measurements. The complex, low-brightness, double-sided jet structure in NGC 1052 makes it a challenge to detect the source at millimeter (mm) wavelengths. However, our first EHT observations have demonstrated that detection is possible up to at least 230 GHz. This study offers a glimpse through the dense surrounding torus and into the innermost central region, where the jets are formed. This has enabled us to finally resolve this region and provide improved constraints on its expansion and magnetic field strength., Comment: 22 pages, 10 figures, published in A&A

Virtual try-on has emerged as a pivotal task at the intersection of computer vision and fashion, aimed at digitally simulating how clothing items fit on the human body. Despite notable progress in single-image virtual try-on (VTO), current methodologies often struggle to preserve a consistent and authentic appearance of clothing across extended video sequences. This challenge arises from the complexities of capturing dynamic human pose and maintaining target clothing characteristics. We leverage pre-existing video foundation models to introduce RealVVT, a photoRealistic Video Virtual Try-on framework tailored to bolster stability and realism within dynamic video contexts. Our methodology encompasses a Clothing & Temporal Consistency strategy, an Agnostic-guided Attention Focus Loss mechanism to ensure spatial consistency, and a Pose-guided Long Video VTO technique adept at handling extended video sequences.Extensive experiments across various datasets confirms that our approach outperforms existing state-of-the-art models in both single-image and video VTO tasks, offering a viable solution for practical applications within the realms of fashion e-commerce and virtual fitting environments., Comment: 10 pages (8 pages main text, 2 pages references), 5 figures in the main text, and 4 pages supplementary materials with 3 additional figures

General-purpose Sparse Matrix-Matrix Multiplication (SpMM) is a fundamental kernel in scientific computing and deep learning. The emergence of new matrix computation units such as Tensor Cores (TCs) brings more opportunities for SpMM acceleration. However, in order to fully unleash the power of hardware performance, systematic optimization is required. In this paper, we propose Acc-SpMM, a high-performance SpMM library on TCs, with multiple optimizations, including data-affinity-based reordering, memory efficient compressed format, high-throughput pipeline, and adaptive sparsity-aware load balancing. In contrast to the state-of-the-art SpMM kernels on various NVIDIA GPU architectures with a diverse range of benchmark matrices, Acc-SpMM achieves significant performance improvements, on average 2.52x (up to 5.11x) speedup on RTX 4090, on average 1.91x (up to 4.68x) speedup on A800, and on average 1.58x (up to 3.60x) speedup on H100 over cuSPARSE., Comment: 11 pages,15 figures, accepted by PPoPP 2025

Affordances, a foundational concept in human-computer interaction and design, have traditionally been explained by direct-perception theories, which assume that individuals perceive action possibilities directly from the environment. However, these theories fall short of explaining how affordances are perceived, learned, refined, or misperceived, and how users choose between multiple affordances in dynamic contexts. This paper introduces a novel affordance theory grounded in Computational Rationality, positing that humans construct internal representations of the world based on bounded sensory inputs. Within these internal models, affordances are inferred through two core mechanisms: feature recognition and hypothetical motion trajectories. Our theory redefines affordance perception as a decision-making process, driven by two components: confidence (the perceived likelihood of successfully executing an action) and predicted utility (the expected value of the outcome). By balancing these factors, individuals make informed decisions about which actions to take. Our theory frames affordances perception as dynamic, continuously learned, and refined through reinforcement and feedback. We validate the theory via thought experiments and demonstrate its applicability across diverse types of affordances (e.g., physical, digital, social). Beyond clarifying and generalizing the understanding of affordances across contexts, our theory serves as a foundation for improving design communication and guiding the development of more adaptive and intuitive systems that evolve with user capabilities., Comment: IUI 2025 Paper

Prostate cancer (PCa) is the most prevalent cancer among men in the United States, accounting for nearly 300,000 cases, 29% of all diagnoses and 35,000 total deaths in 2024. Traditional screening methods such as prostate-specific antigen (PSA) testing and magnetic resonance imaging (MRI) have been pivotal in diagnosis, but have faced limitations in specificity and generalizability. In this paper, we explore the potential of enhancing PCa lesion segmentation using a novel MRI modality called synthetic correlated diffusion imaging (CDI$^s$). We employ several state-of-the-art deep learning models, including U-Net, SegResNet, Swin UNETR, Attention U-Net, and LightM-UNet, to segment PCa lesions from a 200 CDI$^s$ patient cohort. We find that SegResNet achieved superior segmentation performance with a Dice-Sorensen coefficient (DSC) of $76.68 \pm 0.8$. Notably, the Attention U-Net, while slightly less accurate (DSC $74.82 \pm 2.0$), offered a favorable balance between accuracy and computational efficiency. Our findings demonstrate the potential of deep learning models in improving PCa lesion segmentation using CDI$^s$ to enhance PCa management and clinical support., Comment: 8 pages, 2 figures, to be published in Studies in Computational Intelligence. This paper introduces Cancer-Net PCa-Seg, a comprehensive evaluation of deep learning models for prostate cancer segmentation using synthetic correlated diffusion imaging (CDI$^s$). We benchmark five state-of-the-art architectures: U-Net, SegResNet, Swin UNETR, Attention U-Net, and LightM-UNet

This paper proposes a dual divide-and-optimize algorithm (DualOpt) for solving the large-scale traveling salesman problem (TSP). DualOpt combines two complementary strategies to improve both solution quality and computational efficiency. The first strategy is a grid-based divide-and-conquer procedure that partitions the TSP into smaller sub-problems, solving them in parallel and iteratively refining the solution by merging nodes and partial routes. The process continues until only one grid remains, yielding a high-quality initial solution. The second strategy involves a path-based divide-and-optimize procedure that further optimizes the solution by dividing it into sub-paths, optimizing each using a neural solver, and merging them back to progressively improve the overall solution. Extensive experiments conducted on two groups of TSP benchmark instances, including randomly generated instances with up to 100,000 nodes and real-world datasets from TSPLIB, demonstrate the effectiveness of DualOpt. The proposed DualOpt achieves highly competitive results compared to 10 state-of-the-art algorithms in the literature. In particular, DualOpt achieves an improvement gap up to 1.40% for the largest instance TSP100K with a remarkable 104x speed-up over the leading heuristic solver LKH3. Additionally, DualOpt demonstrates strong generalization on TSPLIB benchmarks, confirming its capability to tackle diverse real-world TSP applications., Comment: Accepted by AAAI-25, February 2025

A flow control framework based on linear stability analysis is proposed focusing on reducing the aerodynamic drag due to dynamic stall through a finite-window temporal actuation. The methodology is applied on a periodically plunging SD7003 airfoil.Finite-time Lyapunov exponent (FTLE) fields reveal a saddle point near the airfoil leading edge, where a shear layer forms and feeds a dynamic stall vortex (DSV). A local stability analysis conducted at this saddle point identifies a Kelvin-Helmholtz instability, and the most unstable eigenvalue frequencies remain constant when the variation in the effective angle of attack is minimal. The findings from the FTLE fields and the stability analysis are used to inform the position and finite duty cycle of a periodic blowing and suction actuation applied in a wall-resolved large eddy simulation (LES). The present framework reduces the actuation duty cycle by 77.5% during the airfoil plunging motion, while maintaining the same performance as a continuous actuation throughout the entire cycle. The LES results demonstrate that disturbances from the stability-analysis-informed actuation modify the leading-edge dynamics, preventing the formation of the coherent DSV and significantly reducing the drag., Comment: 23 pages, 10 figures

We introduce MiniMax-01 series, including MiniMax-Text-01 and MiniMax-VL-01, which are comparable to top-tier models while offering superior capabilities in processing longer contexts. The core lies in lightning attention and its efficient scaling. To maximize computational capacity, we integrate it with Mixture of Experts (MoE), creating a model with 32 experts and 456 billion total parameters, of which 45.9 billion are activated for each token. We develop an optimized parallel strategy and highly efficient computation-communication overlap techniques for MoE and lightning attention. This approach enables us to conduct efficient training and inference on models with hundreds of billions of parameters across contexts spanning millions of tokens. The context window of MiniMax-Text-01 can reach up to 1 million tokens during training and extrapolate to 4 million tokens during inference at an affordable cost. Our vision-language model, MiniMax-VL-01 is built through continued training with 512 billion vision-language tokens. Experiments on both standard and in-house benchmarks show that our models match the performance of state-of-the-art models like GPT-4o and Claude-3.5-Sonnet while offering 20-32 times longer context window. We publicly release MiniMax-01 at https://github.com/MiniMax-AI., Comment: A technical report from MiniMax. The authors are listed in alphabetical order. We open-sourced our MiniMax-01 at https://github.com/MiniMax-AI

We present two kinds of resonance soliton solutions on the Ultrahyperbolic space $\mathbb{U}$ for the G=U(2) Yang equation which is equivalent to the anti-self-dual Yang-Mills (ASDYM) equation. We reveal and illustrate the solitonic behaviors in four-dimensional Wess-Zumino-Witten (WZW$_4$) model through the sigma model action densities. The WZW$_4$ model on $\mathbb{U}$ describes a string field theory action of open N=2 string theories and hence our solutions suggest the existence of the corresponding classical objects in the N=2 string theories. Our solutions include multiple-pole solutions and V-shape soliton solutions. The V-shape solitons suggest annihilation and creation process of two solitons and would be building blocks to classify the ASDYM solitons, like the role of Y-shape solitons in classification of the KP (line) solitons. We also clarify the relationship between the Cauchy matrix approach and the binary Darboux transformation in terms of quasideterminants. Our formalism can start with a simpler input data for the soliton solutions and hence might give a suitable framework of the classification of the ASDYM solitons., Comment: 35 pages, 16 figures

Android apps can hold secret strings of themselves such as cloud service credentials or encryption keys. Leakage of such secret strings can induce unprecedented consequences like monetary losses or leakage of user private information. In practice, various security issues were reported because many apps failed to protect their secrets. However, little is known about the types, usages, exploitability, and consequences of app secret leakage issues. While a large body of literature has been devoted to studying user private information leakage, there is no systematic study characterizing app secret leakage issues. How far are Android app secrets from being stolen? To bridge this gap, we conducted the first systematic study to characterize app secret leakage issues in Android apps based on 575 potential app secrets sampled from 14,665 popular Android apps on Google Play. We summarized the common categories of leaked app secrets, assessed their security impacts and disclosed app bad practices in storing app secrets. We devised a text mining strategy using regular expressions and demonstrated that numerous app secrets can be easily stolen, even from the highly popular Android apps on Google. In a follow-up study, we harvested 3,711 distinct exploitable app secrets through automatic analysis. Our findings highlight the prevalence of this problem and call for greater attention to app secret protection., Comment: A pre-print version of the paper. It is accepted to Empirical Software Engineering (EMSE)

The Einstein Probe (EP) is an interdisciplinary mission of time-domain and X-ray astronomy. Equipped with a wide-field lobster-eye X-ray focusing imager, EP will discover cosmic X-ray transients and monitor the X-ray variability of known sources in 0.5-4 keV, at a combination of detecting sensitivity and cadence that is not accessible to the previous and current wide-field monitoring missions. EP can perform quick characterisation of transients or outbursts with a Wolter-I X-ray telescope onboard. In this paper, the science objectives of the Einstein Probe mission are presented. EP is expected to enlarge the sample of previously known or predicted but rare types of transients with a wide range of timescales. Among them, fast extragalactic transients will be surveyed systematically in soft X-rays, which include {\gamma}-ray bursts and their variants, supernova shock breakouts, and the predicted X-ray transients associated with binary neutron star mergers. EP will detect X-ray tidal disruption events and outbursts from active galactic nuclei, possibly at an early phase of the flares for some. EP will monitor the variability and outbursts of X-rays from white dwarfs, neutron stars and black holes in our and neighbouring galaxies at flux levels fainter than those detectable by the current instruments, and is expected to discover new objects. A large sample of stellar X-ray flares will also be detected and characterised. In the era of multi-messenger astronomy, EP has the potential of detecting the possible X-ray counterparts of gravitational wave events, neutrino sources, and ultra-high energy {\gamma}-ray and cosmic ray sources. EP is expected to help advance the studies of extreme objects/phenomena and their underlying physical processes revealed in the dynamic X-ray universe, as well as studies in other areas of X-ray astronomy., Comment: 67 pages, 24 figures, accepted for publication in SCIENCE CHINA Physics, Mechanics & Astronomy