Search

Mass spectrometry analysis of the inner ear fluid, perilymph, cerebrospinal fluid, and vestibular schwannoma proteome and its relation to hearing loss in patients with vestibular schwannomas. Patients with severe to profound hearing loss, compared to those with normal hearing or mild hearing loss, exhibited an upregulation of Complement Factor H-related protein 2 (CFHR2) in the perilymph. This is indicative of increased inflammation in the inner ear, suggesting a potential cause of hearing loss in patients with vestibular schwannomas affecting the ear.

As one of the major causes of antimicrobial resistance, beta-lactamase develops rapidly among bacteria. Detection of beta-lactamase in an efficient and low-cost point-of-care testing (POCT) way is urgently needed. However, due to the volatile environmental factors, the quantitative measurement of current POCT is often inaccurate. Herein, we demonstrate an artificial intelligence (AI)-assisted mobile health system that consists of a paper-based beta-lactamase fluorogenic probe analytical device and a smartphone-based AI cloud. An ultrafast broad-spectrum fluorogenic probe (B1) that could respond to beta-lactamase within 20 s was first synthesized, and the detection limit was determined to be 0.13 nmol/L. Meanwhile, a three-dimensional microfluidic paper-based analytical device was fabricated for integration of B1. Also, a smartphone-based AI cloud was developed to correct errors automatically and output results intelligently. This smart system could calibrate the temperature and pH in the beta-lactamase level detection in complex samples and mice infected with various bacteria, which shows the problem-solving ability in interdisciplinary research, and demonstrates potential clinical benefits., Funding Agencies|National Key R&D Program of China [2020YFA0709900]; National Natural Science Foundation of China [62288102, 22077101, 52073230]; Joint Research Funds of Department of Science & Technology of Shaanxi Province; Northwestern~Polytechnical University [2020GXLH-Z-008, 2020GXLH-Z-013]; Shaanxi Provincial Science Fund for Distinguished Young Scholars [2023-JC-JQ-32]; Key Research and Development Program of Shaanxi [2020ZDLGY13-04]; Fundamental Research Funds for the Central Universities and Innovation Foundation for Doctorate Dissertation of Northwestern Polytechnical University [CX2021121]

Tao W, Li Y, Zhu M, Li C, Li P. Onco Targets Ther. 2019;12:8841–8851. We, the Editor and Publisher of the journal OncoTargets and Therapy have retracted the published article. Following publication of the article, concerns were raised about the duplication of images from Figures 4 and 5 with images from an unrelated article. Specifically, The images for Figure 4D, BGC803 and BGC823, Ctrl, have been duplicated with images for Figure 4A, miR-519a mimics NC, 0Gy and Figure 4G, miR-519a mimics+OE-EphA2, 2Gy, respectively, from Gong S, Li Y, Lv L, et al. Restored microRNA-519a enhances the radiosensitivity of non-small cell lung cancer via suppressing EphA2. Gene Ther. 2022;29:588–600. https://doi.org/10.1038/s41434-020-00213-x. The image for Figure 5D, BGC803, pcNORAD+miR-214, has been duplicated with the image for Figure 4G, si-EphA2, 0Gy from Gong S, et al (2022). The corresponding author did not respond to our queries and was unable to provide a satisfactory explanation for how the images came to be duplicated or provide satisfactory original data for the study. As verifying the validity of published work is core to the integrity of the scholarly record, the Publisher and Editor requested to retract the article and the corresponding author was notified of this. We have been informed in our decision-making by our editorial policies and COPE guidelines. The retracted article will remain online to maintain the scholarly record, but it will be digitally watermarked on each page as “Retracted”.

Metaverse, a hypothetical digital environment linking the cyber world and the physical world, is expected to revolutionize the way people interact. In the metaverse, people interact with objects, the environment, and each other through digital representations of themselves or avatars across time and space. For example, in the metaverse, people can have meetings with colleagues hundreds of miles away. They can also walk through the aisles of a store, find the best fit and have it delivered to their doorstep. It is also possible to simulate the optimal process manufacturing line to adjust for product variation and minimize bottlenecks, or test an innovative aircraft wing design without building expensive prototypes.

Jiaxu Lu,1,2,* Jiaxian Chen,2,* Junhong Ye,2 Zhen Shi,2 Xiang Gao,1 Peicong Chen,2 Yanzhou Chang,2 Hao Lin,2 Peng Li1 1Stem Cell Research and Cellular Therapy Center, Affiliated Hospital of Guangdong Medical University, Zhanjiang, 524001, People’s Republic of China; 2Orthopedic Center, Affiliated Hospital of Guangdong Medical University, Zhanjiang, 524001, People’s Republic of China*These authors contributed equally to this workCorrespondence: Hao Lin, Orthopedic Center, Affiliated Hospital of Guangdong Medical University, Zhanjiang, 524001, People’s Republic of China, Email linhao@gdmu.edu.cn Peng Li, Stem Cell Research and Cellular Therapy Center, Affiliated Hospital of Guangdong Medical University, Zhanjiang, 524001, People’s Republic of China, Email 13763086273@163.comIntroduction: Osteosarcoma is a prevalent and highly malignant primary bone tumor. However, current clinical therapeutic drugs for osteosarcoma are not suitable for long-term use due to significant side effects. Therefore, there is an urgent need to develop new drugs with fewer side effects. Dipsacus asperoides C. Y. Cheng et T. M. Ai, a traditional Chinese medicine, is commonly used for its anti-inflammatory, anti-pain, bone fracture healing, and anti-tumor effects. In this study, we investigated the effects of exosome-like nanoparticles derived from Dipsacus asperoides (DAELNs) on osteosarcoma cells in vitro and in vivo.Methods: DAELNs were isolated and purified from Dipsacus asperoides and their physical and chemical properties were characterized using transmission electron microscopy (TEM) and nanoparticle tracking analysis (NTA). The cellular uptake of DAELNs in osteosarcoma cells was analyzed by PKH26 staining. The proliferation, invasion, migration, and apoptosis of osteosarcoma cells were assessed using CCK8 assay, EdU assay, colony-formation assay, transwell assay, wound healing assay, and mitochondrial membrane potential measurement, respectively. The regul

Microplastics (MPs) is a major threat to agroecosystems. Their accumulation and impacts should be evaluated to advance our understanding of soil function and health. Uncovering the role of cascade effects in regulating crop growth is crucial to understanding the link between MPs disturbance and environmental functions. Therefore, we aimed to assess how the cascade changes between (non-) biological factors and functional traits of maize regulate the response of maize growth to MPs in different nutrient soil environments. We found that soil dehydration induced by MPs may disrupt the balance of the physiological status of maize, negatively affect photosynthetic performance, and enhance competition among organisms for limited nutrients. However, root-responsive nutrient cues with a high degree of tectonic freedom allowed adaptive phenotypic plasticity to occur, masking the negative effects of MPs. In nutrient-rich soil environments, moderate and high intensity (>0.5 %) MPs disturbances initiated root nutrient foraging activities, and maize tended to decrease its cost of investing in root construction, i.e., increasing specific root length (SRL) to promote its own growth. The growth of maize was mainly characterized by increases in the belowground biomass (BGB, 7.11 to 20.81 g) and aboveground biomass (AGB, 61.11 to 118.26 g). Our study suggests that a cascade effect between environmental factors initiated by MPs and the functional architecture of the maize root system drives maize to regulate its growth by responding to nutrient cues. These findings will help to ensure food security, formulate environmental risk management policies and protect soil health, especially in the context of future agriculture.

Improving the efficiency of current neural networks and modeling them in biological neural systems have become popular research directions in recent years. Pulse-coupled neural network (PCNN) is a well applicated model for imitating the computation characteristics of the human brain in computer vision and neural network fields. However, differences between the PCNN and biological neural systems remain: limited neural connection, high computational cost, and lack of stochastic property. In this study, random-coupled neural network (RCNN) is proposed. It overcomes these difficulties in PCNN's neuromorphic computing via a random inactivation process. This process randomly closes some neural connections in the RCNN model, realized by the random inactivation weight matrix of link input. This releases the computational burden of PCNN, making it affordable to achieve vast neural connections. Furthermore, the image and video processing mechanisms of RCNN are researched. It encodes constant stimuli as periodic spike trains and periodic stimuli as chaotic spike trains, the same as biological neural information encoding characteristics. Finally, the RCNN is applicated to image segmentation, fusion, and pulse shape discrimination subtasks. It is demonstrated to be robust, efficient, and highly anti-noised, with outstanding performance in all applications mentioned above.

Language agents have demonstrated autonomous decision-making abilities by reasoning with foundation models. Recently, efforts have been made to train language agents for performance improvement, with multi-step reasoning and action trajectories as the training data. However, collecting such trajectories still requires considerable human effort, by either artificial annotation or implementations of diverse prompting frameworks. In this work, we propose A$^3$T, a framework that enables the Autonomous Annotation of Agent Trajectories in the style of ReAct. The central role is an ActRe prompting agent, which explains the reason for an arbitrary action. When randomly sampling an external action, the ReAct-style agent could query the ActRe agent with the action to obtain its textual rationales. Novel trajectories are then synthesized by prepending the posterior reasoning from ActRe to the sampled action. In this way, the ReAct-style agent executes multiple trajectories for the failed tasks, and selects the successful ones to supplement its failed trajectory for contrastive self-training. Realized by policy gradient methods with binarized rewards, the contrastive self-training with accumulated trajectories facilitates a closed loop for multiple rounds of language agent self-improvement. We conduct experiments using QLoRA fine-tuning with the open-sourced Mistral-7B-Instruct-v0.2. In AlfWorld, the agent trained with A$^3$T obtains a 1-shot success rate of 96%, and 100% success with 4 iterative rounds. In WebShop, the 1-shot performance of the A$^3$T agent matches human average, and 4 rounds of iterative refinement lead to the performance approaching human experts. A$^3$T agents significantly outperform existing techniques, including prompting with GPT-4, advanced agent frameworks, and fully fine-tuned LLMs.

Large Language Models (LLMs) have witnessed remarkable advancements in recent years, prompting the exploration of tool learning, which integrates LLMs with external tools to address diverse real-world challenges. Assessing the capability of LLMs to utilise tools necessitates large-scale and stable benchmarks. However, previous works relied on either hand-crafted online tools with limited scale, or large-scale real online APIs suffering from instability of API status. To address this problem, we introduce StableToolBench, a benchmark evolving from ToolBench, proposing a virtual API server and stable evaluation system. The virtual API server contains a caching system and API simulators which are complementary to alleviate the change in API status. Meanwhile, the stable evaluation system designs solvable pass and win rates using GPT-4 as the automatic evaluator to eliminate the randomness during evaluation. Experimental results demonstrate the stability of StableToolBench, and further discuss the effectiveness of API simulators, the caching system, and the evaluator system.

Tool learning aims to extend the capabilities of large language models (LLMs) with external tools. A major challenge in tool learning is how to support a large number of tools, including unseen tools. To address this challenge, previous studies have proposed retrieving suitable tools for the LLM based on the user query. However, previously proposed methods do not consider the differences between seen and unseen tools, nor do they take the hierarchy of the tool library into account, which may lead to suboptimal performance for tool retrieval. Therefore, to address the aforementioned issues, we propose ToolRerank, an adaptive and hierarchy-aware reranking method for tool retrieval to further refine the retrieval results. Specifically, our proposed ToolRerank includes Adaptive Truncation, which truncates the retrieval results related to seen and unseen tools at different positions, and Hierarchy-Aware Reranking, which makes retrieval results more concentrated for single-tool queries and more diverse for multi-tool queries. Experimental results show that ToolRerank can improve the quality of the retrieval results, leading to better execution results generated by the LLM., Comment: This paper is accepted for LREC-COLING 2024

Large Language Models exhibit robust problem-solving capabilities for diverse tasks. However, most LLM-based agents are designed as specific task solvers with sophisticated prompt engineering, rather than agents capable of learning and evolving through interactions. These task solvers necessitate manually crafted prompts to inform task rules and regulate LLM behaviors, inherently incapacitating to address complex dynamic scenarios e.g., large interactive games. In light of this, we propose Agent-Pro: an LLM-based Agent with Policy-level Reflection and Optimization that can learn a wealth of expertise from interactive experiences and progressively elevate its behavioral policy. Specifically, it involves a dynamic belief generation and reflection process for policy evolution. Rather than action-level reflection, Agent-Pro iteratively reflects on past trajectories and beliefs, fine-tuning its irrational beliefs for a better policy. Moreover, a depth-first search is employed for policy optimization, ensuring continual enhancement in policy payoffs. Agent-Pro is evaluated across two games: Blackjack and Texas Hold'em, outperforming vanilla LLM and specialized models. Our results show Agent-Pro can learn and evolve in complex and dynamic scenes, which also benefits numerous LLM-based applications., Comment: LLM-based Agent

Large Language Models (LLMs) have achieved remarkable performance in objective tasks such as open-domain question answering and mathematical reasoning, which can often be solved through recalling learned factual knowledge or chain-of-thought style reasoning. However, we find that the performance of LLMs in subjective tasks is still unsatisfactory, such as metaphor recognition, dark humor detection, etc. Compared to objective tasks, subjective tasks focus more on interpretation or emotional response rather than a universally accepted reasoning pathway. Based on the characteristics of the tasks and the strong dialogue-generation capabilities of LLMs, we propose RiC (Reasoning in Conversation), a method that focuses on solving subjective tasks through dialogue simulation. The motivation of RiC is to mine useful contextual information by simulating dialogues instead of supplying chain-of-thought style rationales, thereby offering potential useful knowledge behind dialogues for giving the final answers. We evaluate both API-based and open-source LLMs including GPT-4, ChatGPT, and OpenChat across twelve tasks. Experimental results show that RiC can yield significant improvement compared with various baselines.

Recent developments in Multimodal Large Language Models (MLLMs) have shown rapid progress, moving towards the goal of creating versatile MLLMs that understand inputs from various modalities. However, existing methods typically rely on joint training with paired multimodal instruction data, which is resource-intensive and challenging to extend to new modalities. In this paper, we propose a new paradigm through the model composition of existing MLLMs to create a new model that retains the modal understanding capabilities of each original model. Our basic implementation, NaiveMC, demonstrates the effectiveness of this paradigm by reusing modality encoders and merging LLM parameters. Furthermore, we introduce DAMC to address parameter interference and mismatch issues during the merging process, thereby enhancing the model performance. To facilitate research in this area, we propose MCUB, a benchmark for assessing ability of MLLMs to understand inputs from diverse modalities. Experiments on this benchmark and four other multimodal understanding tasks show significant improvements over baselines, proving that model composition can create a versatile model capable of processing inputs from multiple modalities., Comment: Code will be available at https://github.com/THUNLP-MT/ModelCompose

Recent work has made a preliminary attempt to use large language models (LLMs) to solve the stance detection task, showing promising results. However, considering that stance detection usually requires detailed background knowledge, the vanilla reasoning method may neglect the domain knowledge to make a professional and accurate analysis. Thus, there is still room for improvement of LLMs reasoning, especially in leveraging the generation capability of LLMs to simulate specific experts (i.e., multi-agents) to detect the stance. In this paper, different from existing multi-agent works that require detailed descriptions and use fixed experts, we propose a Dynamic Experienced Expert Modeling (DEEM) method which can leverage the generated experienced experts and let LLMs reason in a semi-parametric way, making the experts more generalizable and reliable. Experimental results demonstrate that DEEM consistently achieves the best results on three standard benchmarks, outperforms methods with self-consistency reasoning, and reduces the bias of LLMs., Comment: Accepted by LREC-COLING 2024, Oral presentation

Despite intensive efforts devoted to tool learning, the problem of budget-constrained tool learning, which focuses on resolving user queries within a specific budget constraint, has been widely overlooked. This paper proposes a novel method for budget-constrained tool learning. Our approach involves creating a preferable plan under the budget constraint before utilizing the tools. This plan outlines the feasible tools and the maximum number of times they can be employed, offering a comprehensive overview of the tool learning process for large language models. This allows them to allocate the budget from a broader perspective. To devise the plan without incurring significant extra costs, we suggest initially estimating the usefulness of the candidate tools based on past experience. Subsequently, we employ dynamic programming to formulate the plan. Experimental results demonstrate that our method can be integrated with various tool learning methods, significantly enhancing their effectiveness under strict budget constraints.

Multimodal large language models (MLLMs) have demonstrated promising results in a variety of tasks that combine vision and language. As these models become more integral to research and applications, conducting comprehensive evaluations of their capabilities has grown increasingly important. However, most existing benchmarks fail to consider that, in certain situations, images need to be interpreted within a broader context. In this work, we introduce a new benchmark, named as CODIS, designed to assess the ability of models to use context provided in free-form text to enhance visual comprehension. Our findings indicate that MLLMs consistently fall short of human performance on this benchmark. Further analysis confirms that these models struggle to effectively extract and utilize contextual information to improve their understanding of images. This underscores the pressing need to enhance the ability of MLLMs to comprehend visuals in a context-dependent manner. View our project website at https://thunlp-mt.github.io/CODIS.

While Large language models (LLMs) have demonstrated considerable capabilities across various natural language tasks, they often fall short of the performance achieved by domain-specific state-of-the-art models. One potential approach to enhance domain-specific capabilities of LLMs involves fine-tuning them using corresponding datasets. However, this method can be both resource and time-intensive, and not applicable to closed-source commercial LLMs. In this paper, we propose Preference Adaptation for Enhancing Domain-specific Abilities of LLMs (PANDA), a method designed to augment the domain-specific capabilities of LLMs by leveraging insights from the response preference of expert models without requiring fine-tuning. Our experimental results reveal that PANDA significantly enhances the domain-specific ability of LLMs on text classification and interactive decision tasks. Moreover, LLM with PANDA even outperforms the expert model that being learned on 4 tasks of ScienceWorld. This finding highlights the potential of exploring tuning-free approaches to achieve weak-to-strong generalization.

State-of-the-art Large Multi-Modal Models (LMMs) have demonstrated exceptional capabilities in vision-language tasks. Despite their advanced functionalities, the performances of LMMs are still limited in challenging scenarios that require complex reasoning with multiple levels of visual information. Existing prompting techniques for LMMs focus on either improving textual reasoning or leveraging tools for image preprocessing, lacking a simple and general visual prompting scheme to promote vision-language coordination in LMMs. In this work, we propose Scaffold prompting that scaffolds coordinates to promote vision-language coordination. Specifically, Scaffold overlays a dot matrix within the image as visual information anchors and leverages multi-dimensional coordinates as textual positional references. Extensive experiments on a wide range of challenging vision-language tasks demonstrate the superiority of Scaffold over GPT-4V with the textual CoT prompting. Our code is released in https://github.com/leixy20/Scaffold.

While large language models (LLMs) have been pre-trained on multilingual corpora, their performance still lags behind in most languages compared to a few resource-rich languages. One common approach to mitigate this issue is to translate training data from resource-rich languages into other languages and then continue training. However, using the data obtained solely relying on translation while ignoring the original capabilities of LLMs across languages is not always effective, which we show will limit the performance of cross-lingual knowledge transfer. In this work, we propose SDRRL, a method based on Self-Distillation from Resource-Rich Languages that effectively improve multilingual performance by leveraging the internal capabilities of LLMs on resource-rich languages. We evaluate on different LLMs (LLaMA-2 and SeaLLM) and source languages across various comprehension and generation tasks, experimental results demonstrate that SDRRL can significantly enhance multilingual capabilities while minimizing the impact on original performance in resource-rich languages.

Despite the strong performance of large language models (LLMs) across a wide range of tasks, they still have reliability issues. Previous studies indicate that strong LLMs like GPT-4-turbo excel in evaluating the reliability of responses from LLMs, but face efficiency and local deployment issues. Thus, to enable weak LLMs to effectively assess the reliability of LLM responses, we propose a novel cross-query-comparison-based method called $\textit{Meta Ranking}$ (MR). Unlike previous few-shot methods that solely based on in-context learning capabilities in LLMs, MR assesses reliability by pairwisely ranking the target query-response pair with multiple reference query-response pairs. We found that MR is highly effective in error detection for LLM responses, where weak LLMs, such as Phi-2, could surpass strong baselines like GPT-3.5-turbo, requiring only five reference samples and significantly improving efficiency. We further demonstrate that MR can enhance strong LLMs' performance in two practical applications: model cascading and instruction tuning. In model cascading, we combine open- and closed-source LLMs to achieve performance comparable to GPT-4-turbo with lower costs. In instruction tuning, we use MR for iterative training data filtering, significantly reducing data processing time and enabling LLaMA-7B and Phi-2 to surpass Alpaca-13B with fewer training tokens. These results underscore the high potential of MR in both efficiency and effectiveness., Comment: Preprint, under review. 28 pages

With the bloom of Large Language Models (LLMs), Multimodal Large Language Models (MLLMs) that incorporate LLMs with pre-trained vision models have recently demonstrated impressive performance across diverse vision-language tasks. However, they fall short to comprehend context involving multiple images. A primary reason for this shortcoming is that the visual features for each images are encoded individually by frozen encoders before feeding into the LLM backbone, lacking awareness of other images and the multimodal instructions. We term this issue as prior-LLM modality isolation and propose a two phase paradigm, browse-and-concentrate, to enable in-depth multimodal context fusion prior to feeding the features into LLMs. This paradigm initially "browses" through the inputs for essential insights, and then revisits the inputs to "concentrate" on crucial details, guided by these insights, to achieve a more comprehensive understanding of the multimodal inputs. Additionally, we develop training strategies specifically to enhance the understanding of multi-image inputs. Our method markedly boosts the performance on 7 multi-image scenarios, contributing to increments on average accuracy by 2.13% and 7.60% against strong MLLMs baselines with 3B and 11B LLMs, respectively., Comment: 17 pages, 5 figures

The rapid progress of foundation models has led to the prosperity of autonomous agents, which leverage the universal capabilities of foundation models to conduct reasoning, decision-making, and environmental interaction. However, the efficacy of agents remains limited when operating in intricate, realistic environments. In this work, we introduce the principles of $\mathbf{U}$nified $\mathbf{A}$lignment for $\mathbf{A}$gents ($\mathbf{UA}^2$), which advocate for the simultaneous alignment of agents with human intentions, environmental dynamics, and self-constraints such as the limitation of monetary budgets. From the perspective of $\mathbf{UA}^2$, we review the current agent research and highlight the neglected factors in existing agent benchmarks and method candidates. We also conduct proof-of-concept studies by introducing realistic features to WebShop, including user profiles to demonstrate intentions, personalized reranking for complex environmental dynamics, and runtime cost statistics to reflect self-constraints. We then follow the principles of $\mathbf{UA}^2$ to propose an initial design of our agent, and benchmark its performance with several candidate baselines in the retrofitted WebShop. The extensive experimental results further prove the importance of the principles of $\mathbf{UA}^2$. Our research sheds light on the next steps of autonomous agent research with improved general problem-solving abilities., Comment: Project webpage: https://agent-force.github.io/unified-alignment-for-agents.html

Symbols (or more broadly, non-natural language textual representations) such as numerical sequences, molecular formulas, and table delimiters widely exist, playing important roles in various tasks such as abstract reasoning, chemical property prediction, and table question answering. Despite the impressive natural language comprehension capabilities of large language models (LLMs), their reasoning abilities for symbols remain inadequate, which could attributed to the difference between symbol representations and general natural languages. We propose symbol-to-language (S2L), a tuning-free method that enables large language models to solve symbol-related problems with information expressed in natural language. Specifically, S2L first converts the symbols involved to language-based representations, which can be implemented by prompting LLMs or leveraging external tools, then these language-based representations are integrated into the original problem via direct substitution or concatenation, serving as useful input information for LLMs. We evaluate the S2L method using both API-based (GPT-4, ChatGPT) and open-source (OpenChat) models over eight symbol-related tasks, ranging from symbol-only abstract reasoning to sentiment analysis in social media. Experimental results show that S2L consistently leads to superior performance. For example, by employing S2L for GPT-4, there can be average significant improvements of +21.9% and +9.5% for subtasks in 1D-ARC and Dyck language, respectively. Codes and data are available at https://github.com/THUNLP-MT/symbol2language., Comment: ICLR AGI Workshop 2024

Due to the subjective nature of image quality assessment (IQA), assessing which image has better quality among a sequence of images is more reliable than assigning an absolute mean opinion score for an image. Thus, IQA models are evaluated by global correlation consistency (GCC) metrics like PLCC and SROCC, rather than mean opinion consistency (MOC) metrics like MAE and MSE. However, most existing methods adopt MOC metrics to define their loss functions, due to the infeasible computation of GCC metrics during training. In this work, we construct a novel loss function and network to exploit Global-correlation and Mean-opinion Consistency, forming a GMC-IQA framework. Specifically, we propose a novel GCC loss by defining a pairwise preference-based rank estimation to solve the non-differentiable problem of SROCC and introducing a queue mechanism to reserve previous data to approximate the global results of the whole data. Moreover, we propose a mean-opinion network, which integrates diverse opinion features to alleviate the randomness of weight learning and enhance the model robustness. Experiments indicate that our method outperforms SOTA methods on multiple authentic datasets with higher accuracy and generalization. We also adapt the proposed loss to various networks, which brings better performance and more stable training.

Since the advent of personal computing devices, intelligent personal assistants (IPAs) have been one of the key technologies that researchers and engineers have focused on, aiming to help users efficiently obtain information and execute tasks, and provide users with more intelligent, convenient, and rich interaction experiences. With the development of smartphones and IoT, computing and sensing devices have become ubiquitous, greatly expanding the boundaries of IPAs. However, due to the lack of capabilities such as user intent understanding, task planning, tool using, and personal data management etc., existing IPAs still have limited practicality and scalability. Recently, the emergence of foundation models, represented by large language models (LLMs), brings new opportunities for the development of IPAs. With the powerful semantic understanding and reasoning capabilities, LLM can enable intelligent agents to solve complex problems autonomously. In this paper, we focus on Personal LLM Agents, which are LLM-based agents that are deeply integrated with personal data and personal devices and used for personal assistance. We envision that Personal LLM Agents will become a major software paradigm for end-users in the upcoming era. To realize this vision, we take the first step to discuss several important questions about Personal LLM Agents, including their architecture, capability, efficiency and security. We start by summarizing the key components and design choices in the architecture of Personal LLM Agents, followed by an in-depth analysis of the opinions collected from domain experts. Next, we discuss several key challenges to achieve intelligent, efficient and secure Personal LLM Agents, followed by a comprehensive survey of representative solutions to address these challenges., Comment: https://github.com/MobileLLM/Personal_LLM_Agents_Survey

Rust is a general-purpose programming language designed for performance and safety. Unrecoverable errors (e.g., Divide by Zero) in Rust programs are critical, as they signal bad program states and terminate programs abruptly. Previous work has contributed to utilizing KLEE, a dynamic symbolic test engine, to verify the program would not panic. However, it is difficult for engineers who lack domain expertise to write test code correctly. Besides, the effectiveness of KLEE in finding panics in production Rust code has not been evaluated. We created an approach, called PanicCheck, to hide the complexity of verifying Rust programs with KLEE. Using PanicCheck, engineers only need to annotate the function-to-verify with #[panic_check]. The annotation guides PanicCheck to generate test code, compile the function together with tests, and execute KLEE for verification. After applying PanicCheck to 21 open-source and 2 closed-source projects, we found 61 test inputs that triggered panics; 59 of the 61 panics have been addressed by developers so far. Our research shows promising verification results by KLEE, while revealing technical challenges in using KLEE. Our experience will shed light on future practice and research in program verification.

Real-world scenarios frequently involve multi-objective data-driven optimization problems, characterized by unknown problem coefficients and multiple conflicting objectives. Traditional two-stage methods independently apply a machine learning model to estimate problem coefficients, followed by invoking a solver to tackle the predicted optimization problem. The independent use of optimization solvers and prediction models may lead to suboptimal performance due to mismatches between their objectives. Recent efforts have focused on end-to-end training of predictive models that use decision loss derived from the downstream optimization problem. However, these methods have primarily focused on single-objective optimization problems, thus limiting their applicability. We aim to propose a multi-objective decision-focused approach to address this gap. In order to better align with the inherent properties of multi-objective optimization problems, we propose a set of novel loss functions. These loss functions are designed to capture the discrepancies between predicted and true decision problems, considering solution space, objective space, and decision quality, named landscape loss, Pareto set loss, and decision loss, respectively. Our experimental results demonstrate that our proposed method significantly outperforms traditional two-stage methods and most current decision-focused methods.

Deriving co-speech 3D gestures has seen tremendous progress in virtual avatar animation. Yet, the existing methods often produce stiff and unreasonable gestures with unseen human speech inputs due to the limited 3D speech-gesture data. In this paper, we propose CoCoGesture, a novel framework enabling vivid and diverse gesture synthesis from unseen human speech prompts. Our key insight is built upon the custom-designed pretrain-fintune training paradigm. At the pretraining stage, we aim to formulate a large generalizable gesture diffusion model by learning the abundant postures manifold. Therefore, to alleviate the scarcity of 3D data, we first construct a large-scale co-speech 3D gesture dataset containing more than 40M meshed posture instances across 4.3K speakers, dubbed GES-X. Then, we scale up the large unconditional diffusion model to 1B parameters and pre-train it to be our gesture experts. At the finetune stage, we present the audio ControlNet that incorporates the human voice as condition prompts to guide the gesture generation. Here, we construct the audio ControlNet through a trainable copy of our pre-trained diffusion model. Moreover, we design a novel Mixture-of-Gesture-Experts (MoGE) block to adaptively fuse the audio embedding from the human speech and the gesture features from the pre-trained gesture experts with a routing mechanism. Such an effective manner ensures audio embedding is temporal coordinated with motion features while preserving the vivid and diverse gesture generation. Extensive experiments demonstrate that our proposed CoCoGesture outperforms the state-of-the-art methods on the zero-shot speech-to-gesture generation. The dataset will be publicly available at: https://mattie-e.github.io/GES-X, Comment: The dataset will be released as soon as possible

Restricting bus voltage deviation is crucial for normal operation of multi-bus DC microgrids, yet it has received insufficient attention due to the conflict between two main control objectives in DC microgrids, i.e., voltage regulation and current sharing. By revealing a necessary and sufficient condition for achieving these two objectives, this paper proposes a compromised distributed control algorithm, which regulates the voltage deviation of all buses by relaxing the accuracy of current sharing. Moreover, for a class of DC Microgrids consisting of both critical nodes and ordinary nodes, this paper proposes a distributed control algorithm that restricts the voltage deviation of critical nodes and simultaneously keeps the current sharing of ordinary nodes. This algorithm also works under plug-and-play settings. Simulations illustrate our theory.

In this paper, we introduce Era3D, a novel multiview diffusion method that generates high-resolution multiview images from a single-view image. Despite significant advancements in multiview generation, existing methods still suffer from camera prior mismatch, inefficacy, and low resolution, resulting in poor-quality multiview images. Specifically, these methods assume that the input images should comply with a predefined camera type, e.g. a perspective camera with a fixed focal length, leading to distorted shapes when the assumption fails. Moreover, the full-image or dense multiview attention they employ leads to an exponential explosion of computational complexity as image resolution increases, resulting in prohibitively expensive training costs. To bridge the gap between assumption and reality, Era3D first proposes a diffusion-based camera prediction module to estimate the focal length and elevation of the input image, which allows our method to generate images without shape distortions. Furthermore, a simple but efficient attention layer, named row-wise attention, is used to enforce epipolar priors in the multiview diffusion, facilitating efficient cross-view information fusion. Consequently, compared with state-of-the-art methods, Era3D generates high-quality multiview images with up to a 512*512 resolution while reducing computation complexity by 12x times. Comprehensive experiments demonstrate that Era3D can reconstruct high-quality and detailed 3D meshes from diverse single-view input images, significantly outperforming baseline multiview diffusion methods. Project page: https://penghtyx.github.io/Era3D/.

Cloud-edge collaborative computing paradigm is a promising solution to high-resolution video analytics systems. The key lies in reducing redundant data and managing fluctuating inference workloads effectively. Previous work has focused on extracting regions of interest (RoIs) from videos and transmitting them to the cloud for processing. However, a naive Infrastructure as a Service (IaaS) resource configuration falls short in handling highly fluctuating workloads, leading to violations of Service Level Objectives (SLOs) and inefficient resource utilization. Besides, these methods neglect the potential benefits of RoIs batching to leverage parallel processing. In this work, we introduce Tangram, an efficient serverless cloud-edge video analytics system fully optimized for both communication and computation. Tangram adaptively aligns the RoIs into patches and transmits them to the scheduler in the cloud. The system employs a unique ``stitching'' method to batch the patches with various sizes from the edge cameras. Additionally, we develop an online SLO-aware batching algorithm that judiciously determines the optimal invoking time of the serverless function. Experiments on our prototype reveal that Tangram can reduce bandwidth consumption and computation cost up to 74.30\% and 66.35\%, respectively, while maintaining SLO violations within 5\% and the accuracy loss negligible., Comment: Accepted by IEEE International Conference on Distributed Computing Systems (ICDCS) 2024

Large language models such as ChatGPT and GPT-4 have recently achieved astonishing performance on a variety of natural language processing tasks. In this paper, we propose MANGO, a benchmark to evaluate their capabilities to perform text-based mapping and navigation. Our benchmark includes 53 mazes taken from a suite of textgames: each maze is paired with a walkthrough that visits every location but does not cover all possible paths. The task is question-answering: for each maze, a large language model reads the walkthrough and answers hundreds of mapping and navigation questions such as "How should you go to Attic from West of House?" and "Where are we if we go north and east from Cellar?". Although these questions are easy to humans, it turns out that even GPT-4, the best-to-date language model, performs poorly at answering them. Further, our experiments suggest that a strong mapping and navigation ability would benefit large language models in performing relevant downstream tasks, such as playing textgames. Our MANGO benchmark will facilitate future research on methods that improve the mapping and navigation capabilities of language models. We host our leaderboard, data, code, and evaluation program at https://mango.ttic.edu and https://github.com/oaklight/mango/.

BackgroundHashimoto's thyroiditis (HT) frequently occurs among autoimmune diseases and may simultaneously appear with thyroid cancer. However, it is difficult to diagnose HT at an early stage just by clinical symptoms. Thus, it is urgent to integrate multiple clinical and laboratory factors for the early diagnosis and risk prediction of HT.MethodsWe recruited 1,303 participants, including 866 non-HT controls and 437 diagnosed HT patients. 44 HT patients also had thyroid cancer. Firstly, we compared the difference in thyroid goiter degrees between controls and patients. Secondly, we collected 15 factors and analyzed their significant differences between controls and HT patients, including age, body mass index, gender, history of diabetes, degrees of thyroid goiter, UIC, 25-(OH)D, FT3, FT4, TSH, TAG, TC, FPG, low-density lipoprotein cholesterol, and high-density lipoprotein cholesterol. Thirdly, logistic regression analysis demonstrated the risk factors for HT. For machine learning modeling of HT and thyroid cancer, we conducted the establishment and evaluation of six models in training and test sets.ResultsThe degrees of thyroid goiter were significantly different among controls, HT patients without cancer (HT-C), and HT patients with thyroid cancer (HT+C). Most factors had significant differences between controls and patients. Logistic regression analysis confirmed diabetes, UIC, FT3, and TSH as important risk factors for HT. The AUC scores of XGBoost, LR, SVM, and MLP models indicated appropriate predictive power for HT. The features were arranged by their importance, among which, 25-(OH)D, FT4, and TSH were the top three high-ranking factors.ConclusionsWe firstly analyzed comprehensive factors of HT patients. The proposed machine learning modeling, combined with multiple factors, are efficient for thyroid diagnosis. These discoveries will extensively promote precise diagnosis, personalized therapies, and reduce unnecessary cost for thyroid diseases.

Obstructive Sleep Apnea Syndrome (OSAS) is a common sleep-related disorder. It is characterized by recurrent partial or total collapse of pharyngeal upper airway accompanied by induced vibrations of the soft tissues (e.g., soft palate). The knowledge of the tissue behavior subject to a particular airflow is relevant for realistic clinic applications. However, in-vivo measurements are usually impractical. The goal of the present study is to develop a 3D fluid-structure interaction model for the human uvulopalatal system relevant to OSA based on simplified geometries under physiological conditions. Numerical simulations are performed to assess the influence of the different breathing conditions on the vibrational dynamics of the flexible structure. Meanwhile, the fluid patterns are investigated for the coupled fluid-structure system as well. Increasing the respiratory flow rate is shown to induce larger structural deformation. Vortex shedding induced resonance is not observed due to the large discrepancy between the flow oscillatory frequency and the natural frequency of the structure. The large deformation for symmetric breathing case under intensive respiration is mainly because of the positive feedback from the pressure differences on the top and the bottom surfaces of the structure., QC 20231215

Purpose: This study aims to examine the influence of channel power and influence strategy, in terms of non-coercive strategies, on sanitary equipment manufacturers' relationships with channel members and channel performance. Theoretical framework: The study is based on the literature on channel relationships, which suggests that using an influence strategy can contribute to managing the relationship with the channel members and benefit organization performance. Design/Methodology/Approach: In this study, we sampled from a sanitary equipment manufacturer's channel strategy. We used survey data to examine the effect of channel management strategies from sanitary equipment manufacturers on distributors. Findings: The finding indicates that a supplier using economic power tends to adopt non-coercive strategies. In addition, economic power and non-coercive strategies positively affect the continuity of the relationship with distributors. Relationship continuity between manufacturers and distributors positively impacts whole channel performance. Originality/Value: This study sampled the distributors in the sanitary equipment industry, a market in which the consumers are not knowledgeable about the products. Most consumers base their purchase decisions heavily on the channel member’s recommendations. Therefore, how to manage the relationship with the channel members is critical to understand. Research, Practical & Social implications: Distributors are regarded as an extension of the company's sales capabilities. Channels have always held an essential position in the industry. Maintaining relationships between distributors and improving channel performance is a critical question in distribution management.

Man Li,1,2,* Xintao Li,3,* Hongwei Gao,4 Peng Li,1 Li Zhang,1 Xiaoling Zhang,1 Peipei Liu,5 Xuemei Yang,2 Lili Wu,6 Jiangwei Zeng,2 Shouling Wu,7 Lixia Sun1 1Department of Emergency, The Affiliated Hospital of North China University of Science and Technology, Tangshan, People’s Republic of China; 2Graduate School, North China University of Science and Technology, Tangshan, People’s Republic of China; 3Department of Cardiology, The First Hospital of Soochow University, Jiangsu, People’s Republic of China; 4Department of Emergency, Caofeidian District hospital of Tangshan City, Tangshan, People’s Republic of China; 5School of Public Health, North China University of Science and Technology, Tangshan, People’s Republic of China; 6Department of Cardiology, Shanghai Songjiang District Central Hospital, Shanghai, People’s Republic of China; 7Department of Cardiology, Kailuan Hospital, Tangshan, People’s Republic of China*These authors contributed equally to this workCorrespondence: Lixia Sun, Department of Emergency, The Affiliated Hospital of North China University of Science and Technology, Tangshan, 063000, People’s Republic of China, Email 13323151999@189.cn Shouling Wu, Department of Cardiology, Kailuan General Hospital, Tangshan, 063000, People’s Republic of China, Email drwusl@163.comPurpose: Inflammation plays a critical role in the development of cardiac conduction block (CCB), which is associated with an increased risk of morbidity and mortality. The monocyte-lymphocyte ratio (MLR) acts as a novel inflammatory marker; however, its association with CCB has not yet been studied. This study aimed to investigate the association between MLR and CCB risk.Patients and Methods: In total, 82,472 CCB-free participants were identified from the Kailuan study. MLR was calculated using the monocyte count/lymphocyte count. The participants were stratified based on quartiles of MLR levels. Incident CCB and its subtypes were ascertained f

Objectives We examined the temporal changes of the CSF proteome in patients with herpes simplex encephalitis (HSE) during the course of the disease, in relation to anti-N-methyl-D-aspartate receptor (NMDAR) serostatus, corticosteroid treatment, brain MRI and neurocognitive performance. Methods Patients were retrospectively included from a previous prospective trial with a pre-specified CSF sampling protocol. Mass spectrometry data of the CSF proteome were processed using pathway analysis. Results We included 48 patients (110 CSF samples). Samples were grouped based on time of collection relative to hospital admission – T1: ≤ 9 d, T2: 13–28 d, T3: ≥ 68 d. At T1, a strong multi-pathway response was seen including acute phase response, antimicrobial pattern recognition, glycolysis and gluconeogenesis. At T2, most pathways activated at T1 were no longer significantly different from T3. After correction for multiplicity and considering the effect size threshold, 6 proteins were significantly less abundant in anti-NMDAR seropositive patients compared to seronegative: procathepsin H, heparin cofactor 2, complement factor I, protein AMBP, apolipoprotein A1 and polymeric immunoglobulin receptor. No significant differences in individual protein levels were found in relation to corticosteroid treatment, size of brain MRI lesion or neurocognitive performance. Conclusions We show a temporal change in the CSF proteome in HSE patients during the course of the disease. This study provides insight into quantitative and qualitative aspects of the dynamic pathophysiology and pathway activation patterns in HSE and prompts for future studies on the role of apolipoprotein A1 in HSE, which has previously been associated with NMDAR encephalitis.

In this work, the author theoretically and numerically studies the stimulated Brillouin scattering (SBS) in integrated nano-photonic waveguides. The author first reviews the progress in the field of SBS, including the discoveries of SBS, current progresses in the field of SBS and its applications. Then the author reviews and compares two mechanisms for the SBS in nano-scale waveguides. One of them describes the SBS in waveguides via optical forces induced by electrostriction and optical radiation on the waveguide boundary. The other one is understood from the acoustic wave’s modulation to the optical mode. The modulation is realized via photoelasticity and moving boundary. Both methods are validated by matching the simulation results from author’s models with reported results. In addition, the author analyses and compares the differences and links between these two mechanisms mathematically. To analyse the strength of the SBS (indicated by a parameter called the Brillouin gain), the quality factor (Q factor) of the mechanical mode is of great importance. This is because in order to generate sufficient SBS, it would require the simultaneous confinement of both optical mode and acoustic mode. While optical mode, in most designs, are well confined. The acoustic mode, on the other hand is not well confined and leaky. Q factor, reflecting how much energy is lost, is a crucial parameter in calculating the final SBS gain coefficient. The mechanical Q factor is decided by various resources. Among all the loss factors, anchor loss is believed to be the main loss in high frequency resonators. One method to obtain accurate Q factor is via perfectly matched layer (PML). However. one has to choose proper parameters of the PML to optimize its performance. Based on the mathematical analysis of PML and the well-researched beam resonator structure, the author proposes a novel method for setting up a well-behaved PML. The results show that the choice of a parameter defined as PML sca

With the ever-increasing applications comes the realization that efforts and complexity for developing hardware to keep pace with such compute demands are growing at an even faster rate. And, the problem goes further. As the target cadence of Moore's law is already slipping, more burden is placed on the design methodology to achieve the "equivalent scaling". The proliferation of everywhere machine learning (ML) reveals its multi-faceted role: the killer applications that pull transition to novel hardware and compute paradigms (i.e., system for ML), and the important boosters to design methodology that push toward automated and agile hardware development (i.e., ML for system). Aiming to foster the virtuous cycle between ML and hardware, my research features hardware agile development empowered by ML and studies how to infuse intelligence, improve agility, and eventually enable no-human-in-the-loop automation for scalable and efficacious hardware development flow by synergistic investigation across algorithm, architecture, and electronic design automation (EDA).Specifically, we investigate how different ML techniques can be applied for (1) fast and accurate design evaluation, (2) efficient and scalable design optimization, and (3) high-quality and productive design verification. In design evaluation, we leverage the inherent graph structures of data flow graphs and circuits and explore how domain knowledge can be infused into graph neural network (GNN)-based models, so that we can reconcile timeliness, accuracy, and generalization capability in high-level synthesis (HLS) and logic synthesis performance predictions. In design optimization, we exploit deep reinforcement learning for flexible, scalable, and automated design exploration in HLS resource allocation and workload placement optimization, which is efficient in large search spaces and can be transferred to new designs. In design verification, we utilize the message-passing mechanism in GNN computation to imitate

Anaplastic lymphoma kinase (ALK) positive large B-cell lymphoma (ALK+ LBCL) is an aggressive and rare subtype of B-cell lymphoma. Patients typically present with advanced clinical stage disease and do not respond to conventional chemotherapy; the median overall survival is 1.8 years. The genetic landscape of this entity remains poorly understood. Here we report a unique case of ALK+ LBCL harbouring a rare TFG::ALK fusion. Targeted next-generation sequencing showed no significant single nucleotide variants, insertions/deletions, or other structural variants beyond the TFG::ALK fusion; deep deletions of FOXO1, PRKCA, and the MYB locus were also detected. Our case report draws attention to this rare disease, highlights a need for larger genetic profiling studies, and focuses on the pathogenesis and potential therapeutic targets of this aggressive disease. To our knowledge, this is the first report of a TFG::ALK fusion in ALK+ LBCL.

Development related to PandABox-based fly scans is an important part of the active work on Mamba, the software framework for beamline experiments at the High Energy Photon Source (HEPS); presented in this paper is the progress of our development, and some outlook for advanced fly scans based on knowledge learned during the process. By treating fly scans as a collaboration between a few loosely coupled subsystems - motors / mechanics, detectors / data processing, sequencer devices like PandABox - systematic analyses of issues in fly scans are conducted. Interesting products of these analyses include a general-purpose software-based fly-scan mechanism, a general way to design undulator-monochromator fly scans, a sketch of how to practically implement online tuning of fly-scan behaviours based on processing of the data acquired, and many more. Based on the results above, an architectural discussion on >=10kHz fly scans is given., Comment: 10 pages, 6 figures, accepted by Synchrotron Rad. News

Hong He,1 Peng Liu,2 Peng Li1 1Department of Anesthesiology, Sichuan Provincial People’s Hospital, School of Medicine, University of Electronic Science and Technology of China, Chengdu, Sichuan, 610072 People’s Republic of China; 2Department of Anesthesiology, West China Second University Hospital, Sichuan University, Chengdu, Sichuan, 610044 People’s Republic of ChinaCorrespondence: Peng Li, Department of Anesthesiology, Sichuan Provincial People’s Hospital, School of Medicine,University of Electronic Science and Technology of China, No. 32, West Section 2, 1st Ring Road, Chengdu, Sichuan, 610072, People’s Republic of China, Tel +86 13668169590, Email lipengmazui@163.comObjective: Myocardial ischemia-reperfusion (I/R) injury is a detrimental disease, resulting in high morbidity and mortality globally. In this study, we aimed to investigate the role of Dex in mitigating cardiac I/R injury.Methods: H9c2 cells were treated with Dex (1 μM) for 24 h followed by oxygen-glucose deprivation/re-oxygenation (OGD/R). ANP and BNP mRNA of H9c2 cells and the LDH release were measured. Apoptosis of H9c2 cells was analyzed by flow cytometry. Mitochondrial membrane potential and superoxide production were detected by JC-1 staining and MitoSOXTM Red, respectively. Cell aerobic respiration was measured using Seahorse analysis. In vivo, mice were injected with Dex (25 μg/kg, i.p., once daily) for 5 days and then subjected to heart I/R. Heart function was analyzed by echocardiography. CK-MB and LDH were measured by Elisa. Infarct size was measured using TTC-Evans blue staining. Mitochondrial ultrastructure was observed using transmission electron microscopy. DHE staining, SOD activity, the content of MDA, and the content of GSH/GSSG of heart were measured to evaluate the oxidative stress. In addition, inflammatory cytokines were measured in vivo and in vitro. Furthermore, AMPK, SIRT3, and autophagy-related protein expression in the heart were detected by Western blot.Result