Search

We introduce a method for transferring the signer's appearance in sign language skeletal poses while preserving the sign content. Using estimated poses, we transfer the appearance of one signer to another, maintaining natural movements and transitions. This approach improves pose-based rendering and sign stitching while obfuscating identity. Our experiments show that while the method reduces signer identification accuracy, it slightly harms sign recognition performance, highlighting a tradeoff between privacy and utility. Our code is available at \url{https://github.com/sign-language-processing/pose-anonymization}.

Mother-child interaction is a highly dynamic process neurally characterized by inter-brain synchrony (IBS) at {\theta} and/or {\alpha} rhythms. However, their establishment, dynamic changes, and roles in mother-child interactions remain unknown. Through dynamic analysis of dual-EEG from 40 mother-child dyads during turn-taking cooperation, we uncover that {\theta}-IBS and {\alpha}-IBS alternated with interactive behaviors, with EEG frequency-shift as a prerequisite for IBS transitions. When mothers attempt to track their children's attention and/or predict their intentions, they will adjust their EEG frequencies to align with their children's {\theta} oscillations, leading to a higher occurrence of the {\theta}-IBS state. Conversely, the {\alpha}-IBS state, accompanied by the EEG frequency-shift to the {\alpha} range, is more prominent during mother-led interactions. Further exploratory analysis reveals greater presence and stability of the {\theta}-IBS state during cooperative than non-cooperative conditions, particularly in dyads with stronger emotional attachments and more frequent interactions in their daily lives. Our findings shed light on the neural oscillational substrates underlying the IBS dynamics during mother-child interactions., Comment: 27 Pages,6 figures

The Gamma-Ray Integrated Detectors (GRID) are a space science mission that employs compact gamma-ray detectors mounted on NanoSats in low Earth orbit (LEO) to monitor the transient gamma-ray sky. Owing to the unpredictability of the time and location of gamma-ray bursts (GRBs), obtaining the photon responses of gamma-ray detectors at various incident angles is important for the scientific analysis of GRB data captured by GRID detectors. For this purpose, a dedicated Monte Carlo simulation framework has been developed for GRID detectors. By simulating each GRID detector and the NanoSat carrying it, the spectral energy response, detection efficiency, and other angular responses of each detector for photons with different incident angles and energies can be obtained within this framework. The accuracy of these simulations has been corroborated through on-ground calibration, and the derived angular responses have been successfully applied to the data analysis of recorded GRBs., Comment: 15 pages, 9 figures

With data observed by the Hard X-ray Modulation Telescope (\textit{Insight}-HXMT) and the Neutron star Interior Composition Explorer (\textit {NICER}), we study low-frequency quasi-periodic oscillations (LFQPOs) of the black hole candidate MAXI J1803$-$298 during the 2021 outburst. Based on hardness intensity diagram and difference of the QPOs properties, Type-C and Type-B QPOs are found in the low-hard state and soft intermediate state, respectively. After searching for and classifying QPOs in Fourier domains, we extract the QPO component and study it with wavelet analysis. The QPO and no-QPO time intervals are separated by the confidence level, so that the S-factor, which is defined as the ratio of the QPO time interval to the total length of good time interval, is calculated. We found S-factors decrease with QPOs frequency for Type-C QPOs but stay stable around zero for Type-B QPOs. The relation of S-factor of Type-C QPOs and photon energy, the correlation of S-factor and counts are also studied. Different correlation of S-factor and counts for different energy bands indicates different origins of QPOs in high energy and low energy bands, which may be due to a dual-corona scenario., Comment: 11 pages, 11 figures, 2 tables, MNRAS in press

Dataset Distillation (DD) seeks to create a condensed dataset that, when used to train a model, enables the model to achieve performance similar to that of a model trained on the entire original dataset. It relieves the model training from processing massive data and thus reduces the computation resources, storage, and time costs. This paper illustrates our solution that ranks 1st in the ECCV-2024 Data Distillation Challenge (track 1). Our solution, Modified Difficulty-Aligned Trajectory Matching (M-DATM), introduces two key modifications to the original state-of-the-art method DATM: (1) the soft labels learned by DATM do not achieve one-to-one correspondence with the counterparts generated by the official evaluation script, so we remove the soft labels technique to alleviate such inconsistency; (2) since the removal of soft labels makes it harder for the synthetic dataset to learn late trajectory information, particularly on Tiny ImageNet, we reduce the matching range, allowing the synthetic data to concentrate more on the easier patterns. In the final evaluation, our M-DATM achieved accuracies of 0.4061 and 0.1831 on the CIFAR-100 and Tiny ImageNet datasets, ranking 1st in the Fixed Images Per Class (IPC) Track., Comment: ECCV 2024 Dataset Distillation Challenge

The success of current Entity Alignment (EA) task depends largely on the supervision information provided by labeled data. Considering the cost of labeled data, most supervised methods are difficult to apply in practical scenarios. Therefore, more and more works based on contrastive learning, active learning or other deep learning techniques have been developed, to solve the performance bottleneck caused by the lack of labeled data. However, the existing unsupervised EA methods still have some limitations, either their modeling complexity is high or they cannot balance the effectiveness and practicality of alignment. To overcome these issues, we propose a Simplifying and Learnable graph convolutional attention network for Unsupervised Knowledge Graphs alignment method (SLU). Specifically, we first introduce LCAT, a new and simple framework as the backbone network to model the graph structure of two KGs. Then we design a reconstruction method of relation structure based on potential matching relations for efficiently filtering invalid neighborhood information of aligned entities, to improve the usability and scalability of SLU. Impressively, a similarity function based on consistency is proposed to better measure the similarity of candidate entity pairs. Finally, we conduct extensive experiments on three datasets of different sizes (15K and 100K) and different types (cross-lingual and monolingual) to verify the superiority of SLU. Experimental results show that SLU significantly improves alignment accuracy, outperforming 25 supervised or unsupervised methods, and improving 6.4% in Hits@1 over the best baseline in the best case., Comment: 14 pages, 3 figures

Optimization problems are prevalent across various scenarios. Formulating and then solving optimization problems described by natural language often requires highly specialized human expertise, which could block the widespread application of optimization-based decision making. To make problem formulating and solving automated, leveraging large language models (LLMs) has emerged as a potential way. However, this kind of way suffers from the issue of optimization generalization. Namely, the accuracy of most current LLM-based methods and the generality of optimization problem types that they can model are still limited. In this paper, we propose a unified learning-based framework called LLMOPT to boost optimization generalization. Starting from the natural language descriptions of optimization problems and a pre-trained LLM, LLMOPT constructs the introduced five-element formulation as a universal model for learning to define diverse optimization problem types. Then, LLMOPT employs the multi-instruction tuning to enhance both problem formalization and solver code generation accuracy and generality. After that, to prevent hallucinations in LLMs, such as sacrificing solving accuracy to avoid execution errors, model alignment and self-correction mechanism are adopted in LLMOPT. We evaluate the optimization generalization ability of LLMOPT and compared methods across six real-world datasets covering roughly 20 fields such as health, environment, energy and manufacturing, etc. Extensive experiment results show that LLMOPT is able to model various optimization problem types such as linear/nonlinear programming, mixed integer programming and combinatorial optimization, and achieves a notable 11.08% average solving accuracy improvement compared with the state-of-the-art methods. The code is available at https://github.com/caigaojiang/LLMOPT.

Edge Intelligence (EI) has been instrumental in delivering real-time, localized services by leveraging the computational capabilities of edge networks. The integration of Large Language Models (LLMs) empowers EI to evolve into the next stage: Edge General Intelligence (EGI), enabling more adaptive and versatile applications that require advanced understanding and reasoning capabilities. However, systematic exploration in this area remains insufficient. This survey delineates the distinctions between EGI and traditional EI, categorizing LLM-empowered EGI into three conceptual systems: centralized, hybrid, and decentralized. For each system, we detail the framework designs and review existing implementations. Furthermore, we evaluate the performance and throughput of various Small Language Models (SLMs) that are more suitable for development on edge devices. This survey provides researchers with a comprehensive vision of EGI, offering insights into its vast potential and establishing a foundation for future advancements in this rapidly evolving field.

In recent years, AI-based software engineering has progressed from pre-trained models to advanced agentic workflows, with Software Development Agents representing the next major leap. These agents, capable of reasoning, planning, and interacting with external environments, offer promising solutions to complex software engineering tasks. However, while much research has evaluated code generated by large language models (LLMs), comprehensive studies on agent-generated patches, particularly in real-world settings, are lacking. This study addresses that gap by evaluating 4,892 patches from 10 top-ranked agents on 500 real-world GitHub issues from SWE-Bench Verified, focusing on their impact on code quality. Our analysis shows no single agent dominated, with 170 issues unresolved, indicating room for improvement. Even for patches that passed unit tests and resolved issues, agents made different file and function modifications compared to the gold patches from repository developers, revealing limitations in the benchmark's test case coverage. Most agents maintained code reliability and security, avoiding new bugs or vulnerabilities; while some agents increased code complexity, many reduced code duplication and minimized code smells. Finally, agents performed better on simpler codebases, suggesting that breaking complex tasks into smaller sub-tasks could improve effectiveness. This study provides the first comprehensive evaluation of agent-generated patches on real-world GitHub issues, offering insights to advance AI-driven software development., Comment: 10 pages of main content and 2 pages of references

Reliable responses of service chatbots are often achieved by employing retrieval-based methods that restrict answers to a knowledge base comprising predefined question-answer pairs (QA pairs). To accommodate potential variations in how a customer's query may be expressed, it emerges as the favored solution to augment these QA pairs with similar questions that are possibly diverse while remaining semantic consistency. This augmentation task is known as Similar Question Generation (SQG). Traditional methods that heavily rely on human efforts or rule-based techniques suffer from limited diversity or significant semantic deviation from the source question, only capable of producing a finite number of useful questions. To address these limitations, we propose an SQG approach based on Large Language Models (LLMs), capable of producing a substantial number of diverse questions while maintaining semantic consistency to the source QA pair. This is achieved by leveraging LLMs' natural language understanding capability through fine-tuning with specially designed prompts. The experiments conducted on a real customer-service dataset demonstrate that our method surpasses baseline methods by a significant margin in terms of semantic diversity. Human evaluation further confirms that integrating the answer that reflects the customer's intention is crucial for increasing the number of generated questions that meet business requirements.

This paper proposes a novel metamaterial-inspired absorber designed to enhance the data rate in 52% RF to DC conversion simultaneous wireless information and power transfer system (SWIPT) through biological tissue. The proposed absorber includes split-ring resonators(SRRs) and demonstrates significant permeability characteristics, with both the real and imaginary parts being negative and close to -1. It also improves isolation by around 5dB in a WPT distance of 9mm. A 5mm thick phantom is used for biological tissue in this study. Experimental results exhibits that the SWIPT systems including a rectifier that converts 52% RF to DC efficiency in a WPT distance of 9mm embedding this absorber between power and signal ports at Tx side results in a 5dB improvement in isolation performance. By using proposed absorber, it enables a 7MB/s improvement of data rate and allows signals to be transmitted with 5dBm weaker power than without absorber SWIPT system.

This work considers stepsize schedules for gradient descent on smooth convex objectives. We extend the existing literature and propose a unified technique for constructing stepsizes with analytic bounds for arbitrary iterations. This technique constructs new stepsize schedules by concatenating two short stepsize schedules. Using this approach, we introduce two new families of stepsize schedules, achieving a convergence rate of $O(n^{-\log_2(\sqrt 2+1)})$ with state-of-the-art constants for the objective value and gradient norm of the last iterate, respectively. Furthermore, our analytically derived stepsize schedules either match or surpass the existing best numerically computed stepsize schedules.

In this work, we present UniG, a view-consistent 3D reconstruction and novel view synthesis model that generates a high-fidelity representation of 3D Gaussians from sparse images. Existing 3D Gaussians-based methods usually regress Gaussians per-pixel of each view, create 3D Gaussians per view separately, and merge them through point concatenation. Such a view-independent reconstruction approach often results in a view inconsistency issue, where the predicted positions of the same 3D point from different views may have discrepancies. To address this problem, we develop a DETR (DEtection TRansformer)-like framework, which treats 3D Gaussians as decoder queries and updates their parameters layer by layer by performing multi-view cross-attention (MVDFA) over multiple input images. In this way, multiple views naturally contribute to modeling a unitary representation of 3D Gaussians, thereby making 3D reconstruction more view-consistent. Moreover, as the number of 3D Gaussians used as decoder queries is irrespective of the number of input views, allow an arbitrary number of input images without causing memory explosion. Extensive experiments validate the advantages of our approach, showcasing superior performance over existing methods quantitatively (improving PSNR by 4.2 dB when trained on Objaverse and tested on the GSO benchmark) and qualitatively. The code will be released at https://github.com/jwubz123/UNIG.

Scaling laws offer valuable insights into the design of time series foundation models (TSFMs). However, previous research has largely focused on the scaling laws of TSFMs for in-distribution (ID) data, leaving their out-of-distribution (OOD) scaling behavior and the influence of model architectures less explored. In this work, we examine two common TSFM architectures, encoder-only and decoder-only Transformers, and investigate their scaling behavior on both ID and OOD data. These models are trained and evaluated across varying parameter counts, compute budgets, and dataset sizes. Our experiments reveal that the log-likelihood loss of TSFMs exhibits similar scaling behavior in both OOD and ID settings. We further compare the scaling properties across different architectures, incorporating two state-of-the-art TSFMs as case studies, showing that model architecture plays a significant role in scaling. The encoder-only Transformers demonstrate better scalability than the decoder-only Transformers, while the architectural enhancements in the two advanced TSFMs primarily improve ID performance but reduce OOD scalability. While scaling up TSFMs is expected to drive performance breakthroughs, the lack of a comprehensive understanding of TSFM scaling laws has hindered the development of a robust framework to guide model scaling. We fill this gap in this work by synthesizing our findings and providing practical guidelines for designing and scaling larger TSFMs with enhanced model capabilities.

Large Language Models (LLMs) have been widely used in code completion, and researchers are focusing on scaling up LLMs to improve their accuracy. However, larger LLMs will increase the response time of code completion and decrease the developers' productivity. In this paper, we propose a lightweight and effective LLM for code completion named aiXcoder-7B. Compared to existing LLMs, aiXcoder-7B achieves higher code completion accuracy while having smaller scales (i.e., 7 billion parameters). We attribute the superiority of aiXcoder-7B to three key factors: (1) Multi-objective training. We employ three training objectives, one of which is our proposed Structured Fill-In-the-Middle (SFIM). SFIM considers the syntax structures in code and effectively improves the performance of LLMs for code. (2) Diverse data sampling strategies. They consider inter-file relationships and enhance the capability of LLMs in understanding cross-file contexts. (3) Extensive high-quality data. We establish a rigorous data collection pipeline and consume a total of 1.2 trillion unique tokens for training aiXcoder-7B. This vast volume of data enables aiXcoder-7B to learn a broad distribution of code. We evaluate aiXcoder-7B in five popular code completion benchmarks and a new benchmark collected by this paper. The results show that aiXcoder-7B outperforms the latest six LLMs with similar sizes and even surpasses four larger LLMs (e.g., StarCoder2-15B and CodeLlama-34B), positioning aiXcoder-7B as a lightweight and effective LLM for academia and industry. Finally, we summarize three valuable insights for helping practitioners train the next generations of LLMs for code. aiXcoder-7B has been open-souced and gained significant attention. As of the submission date, aiXcoder-7B has received 2,193 GitHub Stars., Comment: aiXcoder-7B is available at https://github.com/aixcoder-plugin/aiXcoder-7B/tree/main

Current neural audio codecs typically use residual vector quantization (RVQ) to discretize speech signals. However, they often experience codebook collapse, which reduces the effective codebook size and leads to suboptimal performance. To address this problem, we introduce ERVQ, Enhanced Residual Vector Quantization, a novel enhancement strategy for the RVQ framework in neural audio codecs. ERVQ mitigates codebook collapse and boosts codec performance through both intra- and inter-codebook optimization. Intra-codebook optimization incorporates an online clustering strategy and a code balancing loss to ensure balanced and efficient codebook utilization. Inter-codebook optimization improves the diversity of quantized features by minimizing the similarity between successive quantizations. Our experiments show that ERVQ significantly enhances audio codec performance across different models, sampling rates, and bitrates, achieving superior quality and generalization capabilities. It also achieves 100% codebook utilization on one of the most advanced neural audio codecs. Further experiments indicate that audio codecs improved by the ERVQ strategy can improve unified speech-and-text large language models (LLMs). Specifically, there is a notable improvement in the naturalness of generated speech in downstream zero-shot text-to-speech tasks. Audio samples are available here.

Effective research ideation is a critical step for scientific research. However, the exponential increase in scientific literature makes it challenging for researchers to stay current with recent advances and identify meaningful research directions. Recent developments in large language models~(LLMs) suggest a promising avenue for automating the generation of novel research ideas. However, existing methods for idea generation either trivially prompt LLMs or directly expose LLMs to extensive literature without indicating useful information. Inspired by the research process of human researchers, we propose a Chain-of-Ideas~(CoI) agent, an LLM-based agent that organizes relevant literature in a chain structure to effectively mirror the progressive development in a research domain. This organization facilitates LLMs to capture the current advancements in research, thereby enhancing their ideation capabilities. Furthermore, we propose Idea Arena, an evaluation protocol that can comprehensively evaluate idea generation methods from different perspectives, aligning closely with the preferences of human researchers. Experimental results indicate that the CoI agent consistently outperforms other methods and shows comparable quality as humans in research idea generation. Moreover, our CoI agent is budget-friendly, with a minimum cost of \$0.50 to generate a candidate idea and its corresponding experimental design., Comment: 10 pages,5 figures, conference

Large language models excel at performing inference over text to extract information, summarize information, or generate additional text. These inference capabilities are implicated in a variety of ethical harms spanning surveillance, labor displacement, and IP/copyright theft. While many policy, legal, and technical mitigations have been proposed to counteract these harms, these mitigations typically require cooperation from institutions that move slower than technical advances (i.e., governments) or that have few incentives to act to counteract these harms (i.e., the corporations that create and profit from these LLMs). In this paper, we define and build "data defenses" -- a novel strategy that directly empowers data owners to block LLMs from performing inference on their data. We create data defenses by developing a method to automatically generate adversarial prompt injections that, when added to input text, significantly reduce the ability of LLMs to accurately infer personally identifying information about the subject of the input text or to use copyrighted text in inference. We examine the ethics of enabling such direct resistance to LLM inference, and argue that making data defenses that resist and subvert LLMs enables the realization of important values such as data ownership, data sovereignty, and democratic control over AI systems. We verify that our data defenses are cheap and fast to generate, work on the latest commercial and open-source LLMs, resistance to countermeasures, and are robust to several different attack settings. Finally, we consider the security implications of LLM data defenses and outline several future research directions in this area. Our code is available at https://github.com/wagnew3/LLMDataDefenses and a tool for using our defenses to protect text against LLM inference is at https://wagnew3.github.io/LLM-Data-Defenses/.

Generative audio models are rapidly advancing in both capabilities and public utilization -- several powerful generative audio models have readily available open weights, and some tech companies have released high quality generative audio products. Yet, while prior work has enumerated many ethical issues stemming from the data on which generative visual and textual models have been trained, we have little understanding of similar issues with generative audio datasets, including those related to bias, toxicity, and intellectual property. To bridge this gap, we conducted a literature review of hundreds of audio datasets and selected seven of the most prominent to audit in more detail. We found that these datasets are biased against women, contain toxic stereotypes about marginalized communities, and contain significant amounts of copyrighted work. To enable artists to see if they are in popular audio datasets and facilitate exploration of the contents of these datasets, we developed a web tool audio datasets exploration tool at https://audio-audit.vercel.app.

The rapid advancement of Multimodal Large Language Models (MLLMs) has been accompanied by the development of various benchmarks to evaluate their capabilities. However, the true nature of these evaluations and the extent to which they assess multimodal reasoning versus merely leveraging the underlying Large Language Model (LLM) backbone remain unclear. This paper presents a comprehensive investigation into the role of LLM backbones in MLLM evaluation, focusing on two critical aspects: the degree to which current benchmarks truly assess multimodal reasoning and the influence of LLM prior knowledge on performance. Specifically, we introduce a modified evaluation protocol to disentangle the contributions of the LLM backbone from multimodal integration, and an automatic knowledge identification technique for diagnosing whether LLMs equip the necessary knowledge for corresponding multimodal questions. Our study encompasses four diverse MLLM benchmarks and eight state-of-the-art MLLMs. Key findings reveal that some benchmarks allow high performance even without visual inputs and up to 50\% of error rates can be attributed to insufficient world knowledge in the LLM backbone, indicating a heavy reliance on language capabilities. To address knowledge deficiencies, we propose a knowledge augmentation pipeline that achieves significant performance gains, with improvements of up to 60\% on certain datasets, resulting in a approximately 4x increase in performance. Our work provides crucial insights into the role of the LLM backbone in MLLMs, and highlights the need for more nuanced benchmarking approaches.

Incremental Few-Shot Semantic Segmentation (iFSS) tackles a task that requires a model to continually expand its segmentation capability on novel classes using only a few annotated examples. Typical incremental approaches encounter a challenge that the objective of the base training phase (fitting base classes with sufficient instances) does not align with the incremental learning phase (rapidly adapting to new classes with less forgetting). This disconnect can result in suboptimal performance in the incremental setting. This study introduces a meta-learning-based prototype approach that encourages the model to learn how to adapt quickly while preserving previous knowledge. Concretely, we mimic the incremental evaluation protocol during the base training session by sampling a sequence of pseudo-incremental tasks. Each task in the simulated sequence is trained using a meta-objective to enable rapid adaptation without forgetting. To enhance discrimination among class prototypes, we introduce prototype space redistribution learning, which dynamically updates class prototypes to establish optimal inter-prototype boundaries within the prototype space. Extensive experiments on iFSS datasets built upon PASCAL and COCO benchmarks show the advanced performance of the proposed approach, offering valuable insights for addressing iFSS challenges., Comment: conference

In our previous works, we defined Local Information Privacy (LIP) as a context-aware privacy notion and presented the corresponding privacy-preserving mechanism. Then we claim that the mechanism satisfies epsilon-LIP for any epsilon>0 for arbitrary Px. However, this claim is not completely correct. In this document, we provide a correction to the valid range of privacy parameters of our previously proposed LIP mechanism. Further, we propose efficient algorithms to expand the range of valid privacy parameters. Finally, we discuss the impact of updated results on our original paper's experiments, the rationale of the proposed correction and corrected results.

In this paper, we present a new classifier, which integrates significance testing results over different random subspaces to yield consensus p-values for quantifying the uncertainty of classification decision. The null hypothesis is that the test sample has no association with the target class on a randomly chosen subspace, and hence the classification problem can be formulated as a problem of testing for the conjunction of hypotheses. The proposed classifier can be easily deployed for the purpose of conformal prediction and selective classification with reject and refine options by simply thresholding the consensus p-values. The theoretical analysis on the generalization error bound of the proposed classifier is provided and empirical studies on real data sets are conducted as well to demonstrate its effectiveness., Comment: 36 pages, 9 figures

Image fusion is famous as an alternative solution to generate one high-quality image from multiple images in addition to image restoration from a single degraded image. The essence of image fusion is to integrate complementary information from source images. Existing fusion methods struggle with generalization across various tasks and often require labor-intensive designs, in which it is difficult to identify and extract useful information from source images due to the diverse requirements of each fusion task. Additionally, these methods develop highly specialized features for different downstream applications, hindering the adaptation to new and diverse downstream tasks. To address these limitations, we introduce DeFusion++, a novel framework that leverages self-supervised learning (SSL) to enhance the versatility of feature representation for different image fusion tasks. DeFusion++ captures the image fusion task-friendly representations from large-scale data in a self-supervised way, overcoming the constraints of limited fusion datasets. Specifically, we introduce two innovative pretext tasks: common and unique decomposition (CUD) and masked feature modeling (MFM). CUD decomposes source images into abstract common and unique components, while MFM refines these components into robust fused features. Jointly training of these tasks enables DeFusion++ to produce adaptable representations that can effectively extract useful information from various source images, regardless of the fusion task. The resulting fused representations are also highly adaptable for a wide range of downstream tasks, including image segmentation and object detection. DeFusion++ stands out by producing versatile fused representations that can enhance both the quality of image fusion and the effectiveness of downstream high-level vision tasks, simplifying the process with the elegant fusion framework., Comment: 18page

The rapid evolution of Brain-Computer Interfaces (BCIs) has significantly influenced the domain of human-computer interaction, with Steady-State Visual Evoked Potentials (SSVEP) emerging as a notably robust paradigm. This study explores advanced classification techniques leveraging interpretable fuzzy transfer learning (iFuzzyTL) to enhance the adaptability and performance of SSVEP-based systems. Recent efforts have strengthened to reduce calibration requirements through innovative transfer learning approaches, which refine cross-subject generalizability and minimize calibration through strategic application of domain adaptation and few-shot learning strategies. Pioneering developments in deep learning also offer promising enhancements, facilitating robust domain adaptation and significantly improving system responsiveness and accuracy in SSVEP classification. However, these methods often require complex tuning and extensive data, limiting immediate applicability. iFuzzyTL introduces an adaptive framework that combines fuzzy logic principles with neural network architectures, focusing on efficient knowledge transfer and domain adaptation. iFuzzyTL refines input signal processing and classification in a human-interpretable format by integrating fuzzy inference systems and attention mechanisms. This approach bolsters the model's precision and aligns with real-world operational demands by effectively managing the inherent variability and uncertainty of EEG data. The model's efficacy is demonstrated across three datasets: 12JFPM (89.70% accuracy for 1s with an information transfer rate (ITR) of 149.58), Benchmark (85.81% accuracy for 1s with an ITR of 213.99), and eldBETA (76.50% accuracy for 1s with an ITR of 94.63), achieving state-of-the-art results and setting new benchmarks for SSVEP BCI performance.

Best-of-N decoding methods instruct large language models (LLMs) to generate multiple solutions, score each using a scoring function, and select the highest scored as the final answer to mathematical reasoning problems. However, this repeated independent process often leads to the same mistakes, making the selected solution still incorrect. We propose a novel prompting method named Stepwise Correction (StepCo) that helps LLMs identify and revise incorrect steps in their generated reasoning paths. It iterates verification and revision phases that employ a process-supervised verifier. The verify-then-revise process not only improves answer correctness but also reduces token consumption with fewer paths needed to generate. With StepCo, a series of LLMs demonstrate exceptional performance. Notably, using GPT-4o as the backend LLM, StepCo achieves an average accuracy of 94.1 across eight datasets, significantly outperforming the state-of-the-art Best-of-N method by +2.4, while reducing token consumption by 77.8%., Comment: under review

Humanoid robots, with their human-like embodiment, have the potential to integrate seamlessly into human environments. Critical to their coexistence and cooperation with humans is the ability to understand natural language communications and exhibit human-like behaviors. This work focuses on generating diverse whole-body motions for humanoid robots from language descriptions. We leverage human motion priors from extensive human motion datasets to initialize humanoid motions and employ the commonsense reasoning capabilities of Vision Language Models (VLMs) to edit and refine these motions. Our approach demonstrates the capability to produce natural, expressive, and text-aligned humanoid motions, validated through both simulated and real-world experiments. More videos can be found at https://ut-austin-rpl.github.io/Harmon/., Comment: Accepted for oral presentation at 8th Annual Conference on Robot Learning. Project website: https://ut-austin-rpl.github.io/Harmon/

We propose two novel algorithms for constructing convex collision-free polytopes in robot configuration space. Finding these polytopes enables the application of stronger motion-planning frameworks such as trajectory optimization with Graphs of Convex Sets [1] and is currently a major roadblock in the adoption of these approaches. In this paper, we build upon IRIS-NP (Iterative Regional Inflation by Semidefinite & Nonlinear Programming) [2] to significantly improve tunability, runtimes, and scaling to complex environments. IRIS-NP uses nonlinear programming paired with uniform random initialization to find configurations on the boundary of the free configuration space. Our key insight is that finding near-by configuration-space obstacles using sampling is inexpensive and greatly accelerates region generation. We propose two algorithms using such samples to either employ nonlinear programming more efficiently (IRIS-NP2 ) or circumvent it altogether using a massively-parallel zero-order optimization strategy (IRIS-ZO). We also propose a termination condition that controls the probability of exceeding a user-specified permissible fraction-in-collision, eliminating a significant source of tuning difficulty in IRIS-NP. We compare performance across eight robot environments, showing that IRIS-ZO achieves an order-of-magnitude speed advantage over IRIS-NP. IRISNP2, also significantly faster than IRIS-NP, builds larger polytopes using fewer hyperplanes, enabling faster downstream computation. Website: https://sites.google.com/view/fastiris, Comment: 16 pages, 6 figures, accepted for publication in the proceedings of the International Symposium for Robotics Research 2024

Trigonal-PtBi2 has recently garnered significant interest as it exhibits unique superconducting topological surface states due to electron pairing on Fermi arcs connecting bulk Weyl nodes. Furthermore, topological nodal lines have been predicted in trigonal-PtBi2, and their signature was measured in magnetotransport as a dissipationless, i.e. odd under a magnetic field reversal, anomalous planar Hall effect. Understanding the topological superconducting surface state in trigonal-PtBi2 requires unravelling the intrinsic geometric properties of the normal state electronic wavefunctions and further studies of their hallmarks in charge transport characteristics are needed. In this work, we reveal the presence of a strong dissipative, i.e. even under a magnetic field reversal, planar Hall effect in PtBi2 at low magnetic fields and up to room temperature. This robust response can be attributed to the presence of Weyl nodes close to the Fermi energy. While this effect generally follows the theoretical prediction for a planar Hall effect in a Weyl semimetal, we show that it deviates from theoretical expectations at both low fields and high temperatures. We also discuss the origin of the PHE in our material, and the contributions of both the topological features in PtBi2 and its possible trivial origin. Our results strengthen the topological nature of PtBi2 and the strong influence of quantum geometric effects on the electronic transport properties of the low energy normal state., Comment: 7 pages, 5 figures

Large Language Model (LLM)-driven interactive systems currently show potential promise in healthcare domains. Despite their remarkable capabilities, LLMs typically lack personalized recommendations and diagnosis analysis in sophisticated medical applications, causing hallucinations and performance bottlenecks. To address these challenges, this paper proposes MedAide, an LLM-based omni medical multi-agent collaboration framework for specialized healthcare services. Specifically, MedAide first performs query rewriting through retrieval-augmented generation to accomplish accurate medical intent understanding. Immediately, we devise a contextual encoder to obtain intent prototype embeddings, which are used to recognize fine-grained intents by similarity matching. According to the intent relevance, the activated agents collaborate effectively to provide integrated decision analysis. Extensive experiments are conducted on four medical benchmarks with composite intents. Experimental results from automated metrics and expert doctor evaluations show that MedAide outperforms current LLMs and improves their medical proficiency and strategic reasoning., Comment: LLM-based Multi-Agent Collaboration for Medical Applications

In the 6G Internet of Things (IoT) paradigm, unprecedented challenges will be raised to provide massive connectivity, ultra-low latency, and energy efficiency for ultra-dense IoT devices. To address these challenges, we explore the non-orthogonal multiple access (NOMA) based grant-free random access (GFRA) schemes in the cellular uplink to support massive IoT devices with high spectrum efficiency and low access latency. In particular, we focus on optimizing the backoff strategy of each device when transmitting time-sensitive data samples to a multiple-input multiple-output (MIMO)-enabled base station subject to energy constraints. To cope with the dynamic varied channel and the severe uplink interference due to the uncoordinated grant-free access, we formulate the optimization problem as a multi-user non-cooperative dynamic stochastic game (MUN-DSG). To avoid dimensional disaster as the device number grows large, the optimization problem is transformed into a mean field game (MFG), and its Nash equilibrium can be achieved by solving the corresponding Hamilton-Jacobi-Bellman (HJB) and Fokker-Planck-Kolmogorov (FPK) equations. Thus, a Mean Field-based Dynamic Backoff (MFDB) scheme is proposed as the optimal GFRA solution for each device. Extensive simulation has been fulfilled to compare the proposed MFDB with contemporary random access approaches like access class barring (ACB), slotted-Additive Links On-line Hawaii Area (ALOHA), and minimum backoff (MB) under both static and dynamic channels, and the results proved that MFDB can achieve the least access delay and cumulated cost during multiple transmission frames. Keywords: 6G; Internet of Things; grant-free random access; NOMA; dynamic backoff, Comment: 31 pages, 13 figures

We revisit the conjectural method called Schubert analysis for generating the alphabet of symbol letters for Feynman integrals, which was based on geometries of intersecting lines associated with corresponding cut diagrams. We explain the effectiveness of this somewhat mysterious method by relating such geometries to the corresponding Landau singularities, which also amounts to ``uplifting" Landau singularities of a Feynman integral to its symbol letters. We illustrate this {\it Landau-based Schubert analysis} using various multi-loop Feynman integrals in four dimensions and present an automated {\ttfamily Mathematica} notebook for it. We then apply the method to a simplified problem of studying alphabets of physical quantities such as scattering amplitudes and form factors in planar ${\cal N}=4$ super-Yang-Mills. By focusing on a small set of Landau diagrams (as opposed to all relevant Feynman integrals), we show how this method nicely produces the two-loop alphabet of $n$-point MHV amplitudes and that of the $n=4$ MHV form factors. A byproduct of our analysis is an explicit representation of any symbol alphabet obtained this way as the union of various type-$A$ cluster algebras., Comment: 41 pages, 23 figures

Translation is important for cross-language communication, and many efforts have been made to improve its accuracy. However, less investment is conducted in aligning translations with human preferences, such as translation tones or styles. In this paper, a new method is proposed to effectively generate large-scale multilingual parallel corpora with specific translation preferences using Large Language Models (LLMs). Meanwhile, an automatic pipeline is designed to distill human preferences into smaller Machine Translation (MT) models for efficiently and economically supporting large-scale calls in online services. Experiments indicate that the proposed method takes the lead in translation tasks with aligned human preferences by a large margin. Meanwhile, on popular public benchmarks like WMT and Flores, on which our models were not trained, the proposed method also shows a competitive performance compared to SOTA works.

Advancements in reinforcement learning have led to the development of sophisticated models capable of learning complex decision-making tasks. However, efficiently integrating world models with decision transformers remains a challenge. In this paper, we introduce a novel approach that combines the Dreamer algorithm's ability to generate anticipatory trajectories with the adaptive learning strengths of the Online Decision Transformer. Our methodology enables parallel training where Dreamer-produced trajectories enhance the contextual decision-making of the transformer, creating a bidirectional enhancement loop. We empirically demonstrate the efficacy of our approach on a suite of challenging benchmarks, achieving notable improvements in sample efficiency and reward maximization over existing methods. Our results indicate that the proposed integrated framework not only accelerates learning but also showcases robustness in diverse and dynamic scenarios, marking a significant step forward in model-based reinforcement learning.

Fabric-based pneumatic exosuits have a broad application prospect due to their good human-machine interaction performance, but their structural design paradigm has not yet been finalized and requires in-depth research. This paper proposes the concepts of zone inflation and volume transfer for the design of a fabric-based pneumatic exosuit with both efficiency and wearability. The meaning of zone inflation is to divide the inflation area of pneumatic exosuit into inflation-deflation zone and inflation-holding zone which can reduce the consumption of compressed air and improve efficiency. Volume transfer, a strategic distribution method of inflatable regions inside the garment, can effectively enhance the wearability of the exosuit. Using inexpensive thermoplastic polyurethane film and clothing fabric, the exosuit is made by heat pressing and sewing. The exosuit has a response time of 0.5s, a stress area of 1500mm2, and a profile of only 32mm, which can be hidden inside common clothing. A mathematical model is developed to predict the output torque of the exosuit with an error of 3.6%. Mechanical experiments show that the exosuit outputs a torque of 9.1Nm at a pressure of 100kPa. Surface electromyography experiments show that the exosuit can provide users with a boost from sitting to standing, with an average reduction in electromyography signals of 14.95%. The exosuit designed using these methods synthesizes efficiency and wearability and is expected to be an ideal paradigm for fabric-based pneumatic exosuits.

Matrix-variate time series data are increasingly popular in economics, statistics, and environmental studies, among other fields. This paper develops regularized estimation methods for analyzing high-dimensional matrix-variate time series using bilinear matrix-variate autoregressive models. The bilinear autoregressive structure is widely used for matrix-variate time series, as it reduces model complexity while capturing interactions between rows and columns. However, when dealing with large dimensions, the commonly used iterated least-squares method results in numerous estimated parameters, making interpretation difficult. To address this, we propose two regularized estimation methods to further reduce model dimensionality. The first assumes banded autoregressive coefficient matrices, where each data point interacts only with nearby points. A two-step estimation method is used: first, traditional iterated least-squares is applied for initial estimates, followed by a banded iterated least-squares approach. A Bayesian Information Criterion (BIC) is introduced to estimate the bandwidth of the coefficient matrices. The second method assumes sparse autoregressive matrices, applying the LASSO technique for regularization. We derive asymptotic properties for both methods as the dimensions diverge and the sample size $T\rightarrow\infty$. Simulations and real data examples demonstrate the effectiveness of our methods, comparing their forecasting performance against common autoregressive models in the literature.

To enhance resource utilization and address interference issues in ultra-dense networks with mobile edge computing (MEC), a resource utilization approach is first introduced, which integrates orthogonal frequency division multiple access (OFDMA) and non-orthogonal multiple access (NOMA). Then, to minimize the energy consumed by ultra-densely deployed small base stations (SBSs) while ensuring proportional assignment of computational resources and the constraints related to processing delay and security breach cost, the joint optimization of channel selection, the number of subchannels, secure service assignment, multi-step computation offloading, device association, data compression (DC) control, power control, and frequency band partitioning is done for minimizing network-wide energy consumption (EC). Given that the current problem is nonlinear and involves integral optimization parameters, we have devised an adaptive genetic water wave optimization (AGWWO) algorithm by improving the traditional water wave optimization (WWO) algorithm using genetic operations. After that, the computational complexity, convergence, and parallel implementation of AGWWO algorithm are analyzed. Simulation results reveal that this algorithm effectively reduces network-wide EC while guaranteeing the constraints of processing delay and security breach cost.

We perform a theoretical study on the spectrum of the low-lying hidden-bottom ($q\bar{q}b\bar{b}$ with $q=$ $u$, $d$ and $s$) and double-bottom ($q\bar{q}bb$) tetraquark states within a nonrelativistic quark model, in which the instanton-induced interaction is taken as the residual spin-dependent hyperfine interaction between quarks. All the model parameters are fixed by fitting the spectrum of the ground hadron states. The numerical results indicate that masses of several $X_{q\bar{q}}$, $Z_{b}$, and $Z_{bs}$ tetraquark states are below and near thresholds of corresponding meson-meson channels, thus these states may form components of exotic meson states with reasonable probabilities. Especially, in present model, masses of several obtained states with quantum number $I^G(J^P)=1^+(1^{+})$ are close to $Z_b^\pm(10610)$, so one may expect these states to be non-negligible components of the experimentally observed $Z_{b}$ states. Concerning to the double-bottom tetraquark states, the present results are in general consistent with other previous works. Two possible stable $T_{bb}$ states with quark content $bb\bar{n}\bar{n}$ ($n=u$ or $d$ quark) lying at energies $10558$ MeV and $10650$ MeV are found, and one possible stable $bb\bar{n}\bar{s}$ state is found, whose energy is $\sim10687$ MeV., Comment: 15 pages, 7 figures, 12 tables

For predicting software defects at an early stage, researchers have proposed just-in-time defect prediction (JIT-DP) to identify potential defects in code commits. The prevailing approaches train models to represent code changes in history commits and utilize the learned representations to predict the presence of defects in the latest commit. However, existing models merely learn editions in source code, without considering the natural language intentions behind the changes. This limitation hinders their ability to capture deeper semantics. To address this, we introduce a novel bi-modal change pre-training model called BiCC-BERT. BiCC-BERT is pre-trained on a code change corpus to learn bi-modal semantic representations. To incorporate commit messages from the corpus, we design a novel pre-training objective called Replaced Message Identification (RMI), which learns the semantic association between commit messages and code changes. Subsequently, we integrate BiCC-BERT into JIT-DP and propose a new defect prediction approach -- JIT-BiCC. By leveraging the bi-modal representations from BiCC-BERT, JIT-BiCC captures more profound change semantics. We train JIT-BiCC using 27,391 code changes and compare its performance with 8 state-of-the-art JIT-DP approaches. The results demonstrate that JIT-BiCC outperforms all baselines, achieving a 10.8% improvement in F1-score. This highlights its effectiveness in learning the bi-modal semantics for JIT-DP., Comment: Accepted by JSS (The Journal of Systems & Software)

In August of 2024, 495 hackers generated evaluations in an open-ended bug bounty targeting the Open Language Model (OLMo) from The Allen Institute for AI. A vendor panel staffed by representatives of OLMo's safety program adjudicated changes to OLMo's documentation and awarded cash bounties to participants who successfully demonstrated a need for public disclosure clarifying the intent, capacities, and hazards of model deployment. This paper presents a collection of lessons learned, illustrative of flaw reporting best practices intended to reduce the likelihood of incidents and produce safer large language models (LLMs). These include best practices for safety reporting processes, their artifacts, and safety program staffing., Comment: 8 pages, 5 figures

Fragmentation in a gravitationally unstable accretion disk can be an important pathway for forming stellar/planetary companions. To characterize quantitatively the condition for and outcome of fragmentation under realistic thermodynamics, we perform global 3D simulations of gravitationally unstable disks at various cooling rates and cooling types, including the first global simulations of gravitational instability that employ full radiation transport. We find that fragmentation is a stochastic process, with the fragment generation rate per disk area $p_{\rm frag}$ showing an exponential dependence on the parameter $\beta\equiv\Omega_K t_{cool}$, where $\Omega_K$ is the Keplerian rotation frequency and $t_{cool}$ is the average cooling timescale. Compared to a prescribed constant $\beta$, radiative cooling in the optically thin/thick regime makes $p_{\rm frag}$ decrease slower/faster in $\beta$; the critical $\beta$ corresponding to $\sim 1$ fragment per orbit is $\approx$3, 5, 2 for constant $\beta$, optically thin, and optically thick cooling, respectively. The distribution function of the initial fragment mass is remarkably insensitive to disk thermodynamics. Regardless of cooling rate and optical depth, the typical initial fragment mass is $m_{frag} \approx 40 M_{tot}h^3$, with $M_{tot}$ being the total (star+disk) mass and $h=H/R$ being the disk aspect ratio. Applying this result to typical Class 0/I protostellar disks, we find $m_{frag}\sim 20 M_J$, suggesting that fragmentation more likely forms brown dwarfs. Given the finite width of the $m_{frag}$ distribution, forming massive planets is also possible., Comment: 21 pages, 11 figures, accepted for publication in ApJ

3D particle tracking velocimetry (PTV) is a key technique for analyzing turbulent flow, one of the most challenging computational problems of our century. At the core of 3D PTV is the dual-frame fluid motion estimation algorithm, which tracks particles across two consecutive frames. Recently, deep learning-based methods have achieved impressive accuracy in dual-frame fluid motion estimation; however, they heavily depend on large volumes of labeled data. In this paper, we introduce a new method that is completely self-supervised and notably outperforms its fully-supervised counterparts while requiring only 1% of the training samples (without labels) used by previous methods. Our method features a novel zero-divergence loss that is specific to the domain of turbulent flow. Inspired by the success of splat operation in high-dimensional filtering and random fields, we propose a splat-based implementation for this loss which is both efficient and effective. The self-supervised nature of our method naturally supports test-time optimization, leading to the development of a tailored Dynamic Velocimetry Enhancer (DVE) module. We demonstrate that strong cross-domain robustness is achieved through test-time optimization on unseen leave-one-out synthetic domains and real physical/biological domains. Code, data and models are available at https://github.com/Forrest-110/FluidMotionNet., Comment: Accepted by NeurIPS 2024

The composition of pretraining data is a key determinant of foundation models' performance, but there is no standard guideline for allocating a limited computational budget across different data sources. Most current approaches either rely on extensive experiments with smaller models or dynamic data adjustments that also require proxy models, both of which significantly increase the workflow complexity and computational overhead. In this paper, we introduce Adaptive Data Optimization (ADO), an algorithm that optimizes data distributions in an online fashion, concurrent with model training. Unlike existing techniques, ADO does not require external knowledge, proxy models, or modifications to the model update. Instead, ADO uses per-domain scaling laws to estimate the learning potential of each domain during training and adjusts the data mixture accordingly, making it more scalable and easier to integrate. Experiments demonstrate that ADO can achieve comparable or better performance than prior methods while maintaining computational efficiency across different computation scales, offering a practical solution for dynamically adjusting data distribution without sacrificing flexibility or increasing costs. Beyond its practical benefits, ADO also provides a new perspective on data collection strategies via scaling laws., Comment: 21 pages, 10 figures

In this paper, towards the unitary dual problem for classical groups, we propose a refinement of a conjecture of M. Tadi\'c: A representation of good parity is unitary if and only if it is of Arthur type; and a conjectural description of the whole unitary dual. We verify them for representations of corank at most $3$ for symplectic or split odd special orthogonal groups, based on Tadi\'c's classification. To this end, we develop new algorithms to determine whether a given irreducible representation is of Arthur type and give an inductive approach to classify the family of unitary representations that are of Arthur type. As applications, we carry out the explicit list on representations of Arthur type with corank at most $3$, and classify several new families of representations of Arthur type., Comment: Comments are welcome

We study the problem of teaching humanoid robots manipulation skills by imitating from single video demonstrations. We introduce OKAMI, a method that generates a manipulation plan from a single RGB-D video and derives a policy for execution. At the heart of our approach is object-aware retargeting, which enables the humanoid robot to mimic the human motions in an RGB-D video while adjusting to different object locations during deployment. OKAMI uses open-world vision models to identify task-relevant objects and retarget the body motions and hand poses separately. Our experiments show that OKAMI achieves strong generalizations across varying visual and spatial conditions, outperforming the state-of-the-art baseline on open-world imitation from observation. Furthermore, OKAMI rollout trajectories are leveraged to train closed-loop visuomotor policies, which achieve an average success rate of 79.2% without the need for labor-intensive teleoperation. More videos can be found on our website https://ut-austin-rpl.github.io/OKAMI/., Comment: Accepted for oral presentation at 8th Annual Conference on Robot Learning. Project website: https://ut-austin-rpl.github.io/OKAMI/

Deep learning models are increasingly deployed on resource-constrained edge devices for real-time data analytics. In recent years, Vision Transformer models and their variants have demonstrated outstanding performance across various computer vision tasks. However, their high computational demands and inference latency pose significant challenges for model deployment on resource-constraint edge devices. To address this issue, we propose a novel Vision Transformer splitting framework, ED-ViT, designed to execute complex models across multiple edge devices efficiently. Specifically, we partition Vision Transformer models into several sub-models, where each sub-model is tailored to handle a specific subset of data classes. To further minimize computation overhead and inference latency, we introduce a class-wise pruning technique that reduces the size of each sub-model. We conduct extensive experiments on five datasets with three model structures, demonstrating that our approach significantly reduces inference latency on edge devices and achieves a model size reduction of up to 28.9 times and 34.1 times, respectively, while maintaining test accuracy comparable to the original Vision Transformer. Additionally, we compare ED-ViT with two state-of-the-art methods that deploy CNN and SNN models on edge devices, evaluating accuracy, inference time, and overall model size. Our comprehensive evaluation underscores the effectiveness of the proposed ED-ViT framework., Comment: 14 pages, 8 figures

In this paper, we focus on the task of conditional image generation, where an image is synthesized according to user instructions. The critical challenge underpinning this task is ensuring both the fidelity of the generated images and their semantic alignment with the provided conditions. To tackle this issue, previous studies have employed supervised perceptual losses derived from pre-trained models, i.e., reward models, to enforce alignment between the condition and the generated result. However, we observe one inherent shortcoming: considering the diversity of synthesized images, the reward model usually provides inaccurate feedback when encountering newly generated data, which can undermine the training process. To address this limitation, we propose an uncertainty-aware reward modeling, called Ctrl-U, including uncertainty estimation and uncertainty-aware regularization, designed to reduce the adverse effects of imprecise feedback from the reward model. Given the inherent cognitive uncertainty within reward models, even images generated under identical conditions often result in a relatively large discrepancy in reward loss. Inspired by the observation, we explicitly leverage such prediction variance as an uncertainty indicator. Based on the uncertainty estimation, we regularize the model training by adaptively rectifying the reward. In particular, rewards with lower uncertainty receive higher loss weights, while those with higher uncertainty are given reduced weights to allow for larger variability. The proposed uncertainty regularization facilitates reward fine-tuning through consistency construction. Extensive experiments validate the effectiveness of our methodology in improving the controllability and generation quality, as well as its scalability across diverse conditional scenarios. Code will soon be available at https://grenoble-zhang.github.io/Ctrl-U-Page/., Comment: Preprint. Work in progress

Recent approaches for visually-rich document understanding (VrDU) uses manually annotated semantic groups, where a semantic group encompasses all semantically relevant but not obviously grouped words. As OCR tools are unable to automatically identify such grouping, we argue that current VrDU approaches are unrealistic. We thus introduce a new variant of the VrDU task, real-world visually-rich document understanding (ReVrDU), that does not allow for using manually annotated semantic groups. We also propose a new method, ReLayout, compliant with the ReVrDU scenario, which learns to capture semantic grouping through arranging words and bringing the representations of words that belong to the potential same semantic group closer together. Our experimental results demonstrate the performance of existing methods is deteriorated with the ReVrDU task, while ReLayout shows superiour performance.

Closed-source large language models deliver strong performance but have limited downstream customizability. Semi-open models, combining both closed-source and public layers, were introduced to improve customizability. However, parameters in the closed-source layers are found vulnerable to recovery attacks. In this paper, we explore the design of semi-open models with fewer closed-source layers, aiming to increase customizability while ensuring resilience to recovery attacks. We analyze the contribution of closed-source layer to the overall resilience and theoretically prove that in a deep transformer-based model, there exists a transition layer such that even small recovery errors in layers before this layer can lead to recovery failure. Building on this, we propose \textbf{SCARA}, a novel approach that keeps only a few bottom layers as closed-source. SCARA employs a fine-tuning-free metric to estimate the maximum number of layers that can be publicly accessible for customization. We apply it to five models (1.3B to 70B parameters) to construct semi-open models, validating their customizability on six downstream tasks and assessing their resilience against various recovery attacks on sixteen benchmarks. We compare SCARA to baselines and observe that it generally improves downstream customization performance and offers similar resilience with over \textbf{10} times fewer closed-source parameters. We empirically investigate the existence of transition layers, analyze the effectiveness of our scheme and finally discuss its limitations., Comment: 10 pages for main content of the paper

Ti/TiN coatings are utilized in a wide variety of engineering applications due to their superior properties such as high hardness and toughness. Doping Al into Ti/TiN can also enhance properties and lead to even higher performance. Therefore, studying the atomic-level behavior of the TiAl/TiAlN interface is important. However, due to the large number of possible combinations for the 50 mol% Al-doped Ti/TiN system, it is time-consuming to use the DFT-based Monte Carlo method to find the optimal TiAl/TiAlN system with a high work of adhesion. In this study, we use a graph convolutional neural network as an interatomic potential, combined with reinforcement learning, to improve the efficiency of finding optimal structures with a high work of adhesion. By inspecting the features of structures in neural networks, we found that the optimal structures follow a certain pattern of doping Al near the interface. The electronic structure and bonding analysis indicate that the optimal TiAl/TiAlN structures have higher bonding strength. We expect our approach to significantly accelerate the design of advanced ceramic coatings, which can lead to more durable and efficient materials for engineering applications.