Search

Background & Aims Long-term studies of the prognosis of NAFLD are scarce. Here, we investigated the risk of major adverse liver outcomes (MALO) in a large cohort of patients with NAFLD. Methods We conducted a cohort study with data from Swedish university hospitals. Patients (n = 1,260) with NAFLD without cirrhosis were diagnosed through biopsy or radiology, and had fibrosis estimated through vibration-controlled transient elastography, biopsy, or FIB-4 score between 1974 and 2020 and followed up through 2020. Each patient was matched on age, sex, and municipality with up to 10 reference individuals from the general population (n = 12,529). MALO were ascertained from Swedish national registers. The rate of events was estimated by Cox regression. Results MALO occurred in 111 (8.8%, incidence rate = 5.9/1,000 person-years) patients with NAFLD and 197 (1.6%, incidence rate = 1.0/1,000 person-years) reference individuals during a median follow up of 13 years. The rate of MALO was higher in patients with NAFLD (hazard ratio = 6.6; 95% CI = 5.2–8.5). The risk of MALO was highly associated with the stage of fibrosis at diagnosis. In the biopsy subcohort (72% of total sample), there was no difference in risk between patients with and without non-alcoholic steatohepatitis. The 20-year cumulative incidences of MALO were 2% for the reference population, 3% for patients with F0, and 35% for F3. Prognostic information from biopsy was comparable to FIB-4 (C-indices around 0.73 vs. 0.72 at 10 years). Conclusions This study provides updated information on the natural history of NAFLD, showing a high rate of progression to cirrhosis in F3 and a similar prognostic capacity of non-invasive tests to liver biopsy. Impact and implications Several implications for clinical care and future research may be noted based on these results. First, the risk estimates for cirrhosis development are important when communicating risk to patients and deciding on clinical monitoring and treatment.

Heparanase (Hpa1) is expressed by tumor cells and cells of the tumor microenvironment and functions to remodel the extracellular matrix (ECM) and regulate the bioavailability of ECM-bound factors that support tumor growth. Heparanase expression is upregulated in human carcinomas, sarcomas, and hematological malignancies, correlating with increased tumor metastasis, vascular density, and shorter postoperative survival of cancer patients, and encouraging the development of heparanase inhibitors as anti-cancer drugs. Among these are heparin/HS mimetics, the only heparanase-inhibiting compounds that are being evaluated in clinical trials. We have synthesized dicarboxylated oxy-heparins (DCoxHs) containing three carboxylate groups per split residue (DC-Hep). The resulting lead compound (termed XII) was upscaled, characterized, and examined for its effectiveness in tumor models. Potent anti-tumorigenic effects were obtained in models of pancreatic carcinoma, breast cancer, mesothelioma, and myeloma, yielding tumor growth inhibition (TGI) values ranging from 21 to 70% and extending the survival time of the mice. Of particular significance was the inhibition of spontaneous metastasis in an orthotopic model of breast carcinoma following resection of the primary tumor. It appears that apart from inhibition of heparanase enzymatic activity, compound XII reduces the levels of heparanase protein and inhibits its cellular uptake and activation. Heparanase-dependent and -independent effects of XII are being investigated. Collectively, our pre-clinical studies with compound XII strongly justify its examination in cancer patients.

Introduction: Eye emergencies make up nearly 3% of US emergency department (ED) visits. While emergency physicians (EP) should diagnose and treat these ophthalmologic emergencies, many trainees report limited ocular exposure and insufficient training throughout their residency to confidently conduct a thorough slit-lamp exam. Methods: We created an interdisciplinary, simulation-based mastery learning (SBML) curriculum to teach emergency attending physicians how to operate the slit lamp with multimodal learning methodology at a tertiary academic center. The EPs first demonstrate their initial slit-lamp competency with a 20-item checklist, and they then review the necessary curricular content to pass their independent readiness test before completing their in-person teaching and demonstration session with an ophthalmology attending to demonstrate procedural mastery (minimal passing score >90%). Results: Fifteen EPs were enrolled; all completed the final exam of the curriculum. The pre- and post-curriculum checklist scores increased by an average of seven points (P = .002); 86.7% of EPs felt confident in completing a slit-lamp exam after the curriculum, compared to 20% at the beginning. Five of 15 reported teaching learners within the two-month post-curricular period, ranging from 5–30 students. The hands-on teaching was the most positively reviewed element of the curriculum. Conclusion: The SBML program successfully trained EPs on performing a comprehensive slit-lamp exam with promising results of downstream education to junior learners. We encourage other institutions to leverage SBML as a teaching modality for procedural-based training and advocate cross-discipline education initiatives.

In lactating mothers, the high calcium (Ca2+) demand for milk production triggers significant bone loss1. Although oestrogen normally counteracts excessive bone resorption by promoting bone formation, this sex steroid drops precipitously during this postpartum period. Here we report that brain-derived cellular communication network factor 3 (CCN3) secreted from KISS1 neurons of the arcuate nucleus (ARCKISS1) fills this void and functions as a potent osteoanabolic factor to build bone in lactating females. We began by showing that our previously reported female-specific, dense bone phenotype2 originates from a humoral factor that promotes bone mass and acts on skeletal stem cells to increase their frequency and osteochondrogenic potential. This circulatory factor was then identified as CCN3, a brain-derived hormone from ARCKISS1 neurons that is able to stimulate mouse and human skeletal stem cell activity, increase bone remodelling and accelerate fracture repair in young and old mice of both sexes. The role of CCN3 in normal female physiology was revealed after detecting a burst of CCN3 expression in ARCKISS1 neurons coincident with lactation. After reducing CCN3 in ARCKISS1 neurons, lactating mothers lost bone and failed to sustain their progeny when challenged with a low-calcium diet. Our findings establish CCN3 as a potentially new therapeutic osteoanabolic hormone for both sexes and define a new maternal brain hormone for ensuring species survival in mammals.

Qiang-qiang Fan,1 Bing-tao Zhai,1 Dan Zhang,1 Xiao-fei Zhang,1 Jiang-xue Cheng,1 Dong-yan Guo,1– 3 Huan Tian4 1School of Pharmacy, Shaanxi University of Chinese Medicine, Xi’an, 712046, People’s Republic of China; 2Shaanxi Key Laboratory of Chinese Medicine Fundamentals and New Drugs Research, Shaanxi University of Chinese Medicine, Xi’an, 712046, People’s Republic of China; 3State Key Laboratory of Research & Development of Characteristic Qin Medicine Resources (Cultivation), Shaanxi University of Chinese Medicine, Xi’an, 712046, People’s Republic of China; 4Xi’an Hospital of Traditional Chinese Medicine, Xi’an, 712046, People’s Republic of ChinaCorrespondence: Dong-yan Guo; Huan Tian, Email 2051080@sntcm.edu.cn; 18149412681@163.comPurpose: Yinhua Gout Granules (YGG) is a traditional Chinese medicine preparation with a variety of pharmacological effects, and its clinical efficacy in the treatment of gouty arthritis (GA) has been fully confirmed. However, the pharmacodynamic basis of YGG and its anti-inflammatory mechanism of action in GA are unknown. The objective of this study was to identify the active components and molecular mechanisms of YGG in the treatment of GA.Methods: Ultra-performance liquid chromatography-electrospray ionization tandem mass spectrometry (UPLC-ESI-MS/MS) and network pharmacology were used to identify and predict the potential active ingredients and related signaling pathways. Then, we revealed the anti-GA effects of YGG based on pharmacodynamic experiments in GA rats. Finally, we integrated transcriptomics and network pharmacology to elucidate the potential mechanism of action and verified the putative mechanism by molecular docking, immunohistochemical (IHC) and Western blot.Results: We have identified 10 major active components of YGG that may have anti-GA effects, such as ferulic acid, rutin, luteolin, etc. Using molecular docking, we found that 10 major compounds could bind well to TNF, PTGS2, IL-6, IL1β, NOS

Da-Li Xu,1 Xiao-Mei Zhang,2 Xue-Ying Tian,2 Xian-Jun Wang,2 Lin Zhao,1 Meng-Ying Gao,1 Lian-Feng Li,1 Jia-Qi Zhao,1 Wu-Chun Cao,1,3,* Shu-Jun Ding2,* 1Department of Epidemiology, School of Public Health, Cheeloo College of Medicine, Shandong UniversityJinan, Shandong, People’s Republic of China; 2Shandong Provincial Key Laboratory of Communicable Disease Control and Prevention, Department of Communicable Disease Control and Prevention, Shandong Center for Disease Control and Prevention, Jinan, Shandong Province, People’s Republic of China; 3State Key Laboratory of Pathogen and Biosecurity, Beijing Institute of Microbiology and Epidemiology, Beijing, People’s Republic of China*These authors contributed equally to this workCorrespondence: Shu-Jun Ding, Shandong Provincial Key Laboratory of Communicable Disease Control and Prevention, Department of Communicable Disease Control and Prevention, Shandong Center for Disease Control and Prevention, 16992 Jingshi Road, Jinan, Shandong Province, People’s Republic of China, Tel/Fax +86 531 82679587, Email dsj_jn@126.com Wu-Chun Cao, State Key Laboratory of Pathogen and Biosecurity, Beijing Institute of Microbiology and Epidemiology, 20 Dong-da Street, Beijing, People’s Republic of China, Email caowuchun@126.comPurpose: To characterize the cytokine profile of patients with severe fever with thrombocytopenia syndrome (SFTS) in relation to disease severity.Patients and Methods: 60 laboratory-confirmed SFTS patients and 12 healthy individuals from multi-centers in Shandong Province of China were included, and all patients were divided into fatal patients (9) and recovered patients (51) due to their final outcomes. Multiplex-microbead immunoassays were conducted to estimate levels of 27 cytokines in the sera of patients and controls.Results: The results showed that levels of IL-2, IL-4, IL-6, IL-7, IL-8, IL-15, IL-1RA, G-CSF, GM-CSF, IFN-γ, TNF-α, basic FGF, PDGF-BB, RANTES, IP-10, MIP-1α, MIP-1

Groen is een zeer effectief middel om de stad meer klimaatbestendig te maken. Het effect van groen is echter afhankelijk van het type groen en de omgeving waarin dat groen staat. Concrete richtlijnen voor effectief groen in verschillende wijktypen ontbreken tot nog toe. Om in die lacune te voorzien ging in 2019 onder de titel “Effectief groen voor klimaatadaptatie in de stad” een topsector project (TU 18044) van start. In een samenwerking van organisaties uit de groene sector, onderwijs en onderzoek is met subsidie van het ministerie van LNV gewerkt aan het ontwikkelen van concrete handvatten voor ontwerp, aanleg, inrichting en beheer van groen dat effectief kan bijdragen aan klimaatadaptatie in specifieke ruimtelijke situaties in de stad. Daarvoor is bestaande wetenschappelijke kennis vertaald naar richtlijnen en concepten (sjablonen) voor effectief klimaatgroen en zijn metingen en simulaties uitgevoerd om het effect van specifieke groene interventies in standaardsituaties in het stedelijk gebied te kwantificeren. De nu voorliggende brochure presenteert de resultaten voor een viertal in Nederland zeer algemene “wijktypen”: Bloemkoolwijk, Stedelijk Bouwblok, Volkswijk en Verbindingswegen. In de vorm van een schematische visuele weergave (“sjabloon”) wordt aangegeven hoe groen benut kan worden om het optreden van hittestress te beperken. Daarbij is tevens voor elk van de groene interventies aangegeven hoe groot de te behalen effecten m.b.t. reductie van hittestress op hete dagen zijn.

The decarbonization of the construction industry has emerged as a pressing global priority, and the electric machine is the driving force towards achieving this goal. This study focuses on developing an operation scheduling optimization methodology for electric ready-mixed concrete vehicles (ERVs), which is an innovative and environmental solution for the construction industry. Firstly, we propose a systematic problem definition for the ERV operation process that takes into account the unique properties of electric vehicles and RMC delivery tasks. Subsequently, the entire process is formulated as a Markov decision problem (MDP), which facilitates efficient sequential decision-making. Finally, we develop an improved model-based reinforcement learning method, named parallel-masked-decaying Monte Carlo Tree Search (PMD-MCTS), to solve the MDP efficiently. The entire system is validated in a real case study, and the performance of PMD-MCTS is compared with state-of-the-art benchmarks. The results have shown that the proposed MDP formulation is well-suited for addressing RMC delivery tasks. The PMD-MCTS algorithm and one of its ablation algorithms (PM-MCTS) have outperformed the other benchmarks in terms of either minimizing costs or delay, and PMD-MCTS consumes 32% less computation time than PM-MCTS. © 2024 Elsevier B.V.

Hierarchical Multi-Label Classification (HMLC) is a well-established problem that aims at assigning data instances to multiple classes stored in a hierarchical structure. Despite its importance, existing approaches often face two key limitations: (i) They employ dense networks to solely explore the class hierarchy as hard criterion for maintaining taxonomic consistency among predicted classes, yet without leveraging rich semantic relationships between instances and classes; (ii) They struggle to generalize in settings with deep class levels, since the mini-batches uniformly sampled from different levels ignore the varying complexities of data and result in a non-smooth model adaptation to sparse data. To mitigate these issues, we present a Self-Paced Unified Representation (SPUR) learning framework, which focuses on the interplay between instance and classes to flexibly organize the training process of HMLC algorithms. Our framework consists of two lightweight encoders designed to capture the semantics of input features and the topological information of the class hierarchy. These encoders generate unified embeddings of instances and class hierarchy, which enable SPUR to exploit semantic dependencies between them and produce predictions in line with taxonomic constraints. Furthermore, we introduce a dynamic hardness measurement strategy that considers both class hierarchy and instance features to estimate the learning difficulty of each instance. This strategy is achieved by incorporating the propagation loss obtained at each hierarchical level, allowing for a more comprehensive assessment of learning complexity. Extensive experiments on several empirical benchmarks demonstrate the effectiveness and efficiency of SPUR compared to state-of-the-art methods, especially in scenarios with missing features. Copyright © 2024, Association for the Advancement of Artificial Intelligence (www.aaai.org). All rights reserved.

In this short paper, we study Majorana fermions in terms of a modified Chandrasekhar's separation for the Dirac equation in Kerr-Newman-type spacetimes. If the electric charge or magnetic charge is nonzero, we show that nontrivial differentiable time-periodic Majorana fermions do not exist outside the event horizon in Kerr-Newman and Kerr-Newman-AdS spacetimes, and between the event horizon and the cosmological horizon in Kerr-Newman-dS spacetime., Comment: 8 pages

Yin-Gen Luo,&ast; Xiao-Wu Zhang,&ast; He Zhao, Jin-Gui Li, Jiay-Wei Tsauo, Tao Gong, Ai-Xin Ou, Tian-Hao Cong, Wen-Di Kang, Xiao Li Department of Interventional Therapy, National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, People’s Republic of China&ast;These authors contributed equally to this workCorrespondence: Xiao Li; He Zhao, Department of Interventional Therapy, National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, 100021, People’s Republic of China, Tel/Fax +86 010 8778 8502, Email simonlixiao@gmail.com; hezhaoo@gmail.comObjective: This study aimed to establish a rat model that simulates benign esophageal strictures induced by endoscopic submucosal dissection (ESD).Materials and Methods: Sixteen male Sprague-Dawley rats were randomly divided into mucosal resection (n = 8) and sham-operated groups (n = 8). The rats in the mucosal resection group underwent a 5-mm three-fourths mucosal resection by way of a 3-mm incision in the distal esophagus under direct visualization via laparotomy. Rats in the sham-operated group underwent a 3-mm incision of the muscularis propria layer in the distal esophagus via laparotomy without mucosal resection. Dysphagia score, weight gain, mucosal constriction rate, and histology were evaluated 2 weeks after surgery.Results: Technical success was achieved in all the animals. One rat in the mucosal resection group died of infection, and no other complications were observed. Weight gain (P < 0.001) and luminal diameter derived from the esophagograms (P < 0.001) were significantly lower in the mucosal resection group than those in the sham-operated group. Dysphagia score (P < 0.001) and mucosal constriction rate (P < 0.001) were significantly higher in the mucosal resection group than those

The development of modern cities has often led to increased traffic congestion and limited usable space. One effective solution to these problems is to construct roadway tunnels, which can expand urban space and alleviate traffic congestion. However, building large underpass tunnels in urban areas, especially in soft ground, presents technical challenges owing to its extensive environmental disturbance. The pipe-umbrella box jacking method offers a promising solution to these challenges as this method does not require breaking the ground from the surface or disrupting surface traffic. This study reviews the key techniques of the pipe-umbrella box jacking method for constructing large rectangular tunnels with small burial depths in soft ground. The study focuses on three main technical issues: pipe umbrella installation, soil face excavation, and box jacking control. Furthermore, two tunnel projects in Shanghai, China, are presented as case studies to showcase the construction challenges and countermeasures during box jacking tunneling with pipe-umbrella. Practical construction experience has revealed that the installation of the pipe umbrella before box jacking serves as an effective method of ground pre-support. However, its performance can be influenced by factors such as pipe installation deviation and inter-pipe connection design. Both open-face steel grid extrusion (SGE) shields and closed-face earth pressure balanced (EPB) shields can be used for soil excavation during box jacking, with different work modes resulting in distinct ground deformation patterns. Additionally, lubrication grouting has been found to effectively reduce friction resistance and affect vertical ground deformation. The study also examines the lessons learned from case studies and proposes optimization measures. This study can serve as a valuable reference for the tunneling industry, providing insights into the pipe-umbrella box jacking method and its application in constructing large r, Geo-engineering

A time-domain inverse aeroacoustic method based on the convective Ffowcs Williams–Hawkings equation is presented. The method allows to determine, in real-time, the unsteady forces exerted on rotating blades in the presence of a moving medium. The inversion procedure is based on a space-time regularization with a mixed l1,2-norm, which guarantees accuracy and smoothness of the solution. The method is initially verified through synthetic acoustic signals emitted by rotating sources in a constant flow, up to a convective Mach number of about 0.88. Then the method is validated through signals generated by a propeller immersed in a wind-tunnel jet flow, up to a Mach number of 0.06. Due to the reduced convective Mach number, the leading aeroacoustic effect is derived from a variation of the blade loading. It is argued that the onset of flow separation at high values of the rotor advance ratio is responsible for the onset of force fluctuations that the inverse method is able to retrieve both qualitatively and quantitatively., Wind Energy

The Higgs boson pair production at the LHC provides a probe to the Higgs boson self-coupling. The higher-order QCD corrections in this process are sizable and must be taken into account in comparison with data. Due to the small cross section, it is necessary to consider at least one of the Higgs bosons decaying to bottom quarks. The QCD corrections to the decay processes would also be important in such cases. We present a full calculation of the total and differential cross sections for the $b\bar{b}\gamma\gamma$ final state with next-to-leading order (NLO) QCD corrections. After applying typical kinematic cuts in the final state, we find that QCD NLO corrections in the decay decrease the LO result by $19\%$ and reduce the scale uncertainties by a factor of two. The QCD corrections to the invariant mass $m_{jj\gamma\gamma}$ distribution, the transverse momentum spectra of the leading bottom quark jet and photon are significant and can not be approximated by a constant factor., Comment: 19 pages, 4 figures

Laminaria japonica (LJ) is a common seaweed, which has been widely used as functional food and medicine, but its pharmaceutical activities in angiogenic regulation are unclear. Here, different lines of studies provide evidence to the suppressive effects of LJ extract in VEGF-medicated angiogenic function. In cellular level, LJ extract significantly inhibited VEGF-mediated endothelial cell proliferation, migration, invasion and tube formation. For in vivo verification, LJ extract suppressed the VEGF-induced formation of sub-intestinal vessels in zebrafish embryo model. In signalling cascades, LJ extract decreased the amount of phosphorylated VEGF receptor-2, as triggered by VEGF, and subsequently some key related downstream elements were obviously inactivated. In parallel, NO production and ROS formation, as induced by VEGF, were significantly attenuated. Thus, these preclinical findings promoted the possible development of LJ extract as a potential therapeutic agent for prevention and clinical treatment of angiogenesis-related diseases. © 2024 The Authors

Robotics holds great potential to improve productivity in construction, and coverage path planning (CPP) is an essential capability crucial to various applications, including floor cleaning and environmental monitoring. However, there is still a lack of a comprehensive CPP system that can handle complex indoor conditions and various robotic properties. An improved CPP (ICCP) system that leverages building information modeling (BIM) and robotic configurations for indoor robots is proposed in this study. Firstly, BIM is semantically enriched to generate semantic trapezoidal grid maps (TGMs); Next, a novel concept called "coverage bonus" is incorporated into coverage pattern analysis to enable farsighted decision making; Finally, the coverage sequence is optimized by solving the cluster generalized traveling salesman problem, resulting in routes that minimize both coverage distances and disruptions in indoor activities. Experimental validation shows that the ICPP system can not only attain optimal coverage performance with the highest coverage ratio (97.6%) but also ensure the adherence to indoor coverage rules. Future research will focus on enhancing coverage ratio through tunable hyperparameters, optimizing computation time in ICCP, and expanding the study to multi-robot scenarios.

Silicon-based materials still face critical issues including large volume expansion and poor electrical conductivity during the charging and discharging process. Despite the carbon coating strategy can improve performance, the practical operation of silicon-carbon anodes still suffers from undesirable cycling performance. Herein, we propose a novel method to construct Si/SiOx@G@C microparticles, where the Si/SiOx core is coated by conducive and flexible amorphous carbon networks and graphite shells. The Si/SiOx@G@C demonstrated a high discharge capacity of 821 mA h g−1 and an initial Coulombe efficiency of 82.5 % under the current density of 0.1A/g. At a current density of 1.0 A/g, the Si/SiOx@G@C electrodes delivered a high rechargeable capacity retention of 75.1 % and only 117 % volume expansion over 1,000 charge–discharge cycles. The demonstration of silicon-based batteries with a long cycle life is an important step towards the development of this field beyond Li-ion technology. © 2023

Qi Kong,1 Ling-ling Qian,1 Lei Zhang,1 Huan-huan Liu,1,2 Fan Yang,1 Xiao-lu Zhang,1 Chao Wang,1 Xiao-xi Zhao,1 Ku-lin Li,1 Ru-xing Wang1,2 1Department of Cardiology, The Affiliated Wuxi People’s Hospital of Nanjing Medical University, Wuxi People’s Hospital, Wuxi Medical Center, Nanjing Medical University, Wuxi, 214023, People’s Republic of China; 2Wuxi School of Medicine, Jiangnan University, Wuxi, 214122, People’s Republic of ChinaCorrespondence: Ru-xing Wang; Ku-lin Li, Department of Cardiology, The Affiliated Wuxi People’s Hospital of Nanjing Medical University, Wuxi People’s Hospital, Wuxi Medical Center, Nanjing Medical University, Wuxi, 214023, People’s Republic of China, Email ruxingw@aliyun.com; liyantong2002@126.comPurpose: The aim of this study was to investigate the effects and mechanisms of SGLT2 inhibitor empagliflozin on diabetic coronary function.Methods: A rat diabetic model was established by injection of streptozotocin. Rats in the treated group were administered empagliflozin by gavage and rat coronary vascular tensions were measured after eight weeks. Large conductance calcium activated K+ channel currents were recorded using a patch clamp technique, while human coronary artery smooth muscle cells were used to explore the underlying mechanisms.Results: After incubation with empagliflozin (10, 30, 100, 300, 1000 μmol/L), the Δ relaxation % of rat coronary arteries were 2.459 ± 1.304, 3.251 ± 1.119, 6.946 ± 3.407, 28.36 ± 11.47, 86.90 ± 3.868, respectively. Without and with empagliflozin in the bath solution, BK channel opening probabilities at a membrane potential of +60 mV were 0.0458 ± 0.0517 and 0.3413 ± 0.2047, respectively (p < 0.05, n = 4 cells). After incubation with iberiotoxin, the Δ tensions of rat coronary arteries in the control (Ctrl), untreated (DM), low empagliflozin (10 mg/kg/d)-treated (DM+L-EMPA) and high empagliflozin (30mg/kg/d)-treated (DM+H-EMPA) group were 103.20 ± 5.85, 40.37 ± 22.12, 99.47 ±

Next point-of-interest (POI) recommendation has been a hot research topic, which enables new paradigms for kinds of location-based services in real-world scenarios. Due to the privacy concerns and rigorous data regulations, federated learning provides a distributed learning framework to collaboratively train the recommendation model without sharing the highly sensitive POI data with others. However, there exist two main challenges, namely location noise, and balance between personalization and knowledge sharing, seriously restrict the development of the federated next POI recommendation. To this end, in this work, we propose an adaptive clustering based personalized federated learning framework for next POI recommendation with location noise, named CPF-POI, to address the above challenges. In detail, within the local client, a location recovery module can efficiently remove noises under the given assumption from the noisy POI data in which the recovery error bound can be theoretically proved. Then, within the parameter server, an adaptive clustering scheme is proposed to capture the internal relatedness among all clients to augment positive knowledge sharing. In order to make a balance between personalization and knowledge sharing under personalized federated learning framework, we design an alternative optimization process between clustering similar clients and minimizing local personalized loss functions. Finally, extensive experiments are conducted on two diverse real-world datasets to show the advantages of CPF-POI over state-of-the-art methods. improvement across all metrics on average. IEEE

DNA cytosine methylation plays a vital role in repressing retrotransposons, and such derepression is linked with developmental failure, tumorigenesis and aging. DNA methylation patterns are formed by precisely regulated actions of DNA methylation writers (DNA methyltransferases) and erasers (TET, ten-eleven translocation dioxygenases). However, the mechanisms underlying target-specific oxidation of 5mC by TET dioxygenases remain largely unexplored. Here we show that a large low-complexity domain (LCD), located in the catalytic part of Tet enzymes, negatively regulates the dioxygenase activity. Recombinant Tet3 lacking LCD is shown to be hyperactive in converting 5mC into oxidized species in vitro. Endogenous expression of the hyperactive Tet3 mutant in mouse oocytes results in genome-wide 5mC oxidation. Notably, the occurrence of aberrant 5mC oxidation correlates with a consequent loss of the repressive histone mark H3K9me3 at ERVK retrotransposons. The erosion of both 5mC and H3K9me3 causes ERVK derepression along with upregulation of their neighboring genes, potentially leading to the impairment of oocyte development. These findings suggest that Tet dioxygenases use an intrinsic auto-regulatory mechanism to tightly regulate their enzymatic activity, thus achieving spatiotemporal specificity of methylome reprogramming, and highlight the importance of methylome integrity for development. Here the authors show that TET dioxygenases, the erasers of DNA methylation, use a self-limiting mechanism via their LCD domain to ensure adaptable methylome status and protect the genome from excessive oxidative methylation.

Jiangyan Yong,1,&ast; Hongli Shu,1,&ast; Xiao Zhang,2,&ast; Kun Yang,3 Guining Luo,3 Lu Yu,3 Jiaqi Li,3 Hong Huang4 1Department of Clinical Laboratory, Hospital of Chengdu University of Traditional Chinese Medicine, Chengdu, Sichuan, 610072, People’s Republic of China; 2Department of Clinical Laboratory, Chengdu Children Special Hospital, Chengdu, Sichuan, 610031, People’s Republic of China; 3Chengdu University of Traditional Chinese Medicine, Chengdu, Sichuan, 611137, People’s Republic of China; 4Department of Clinical Laboratory, the People’s Hospital of Chongqing Liang Jiang New Area, Chongqing, 401121, People’s Republic of China&ast;These authors contributed equally to this workCorrespondence: Hong Huang, Email hh1643689264@163.comAbstract: Given the unique physiological and pathological characteristics of the lung, the direct, inhalable route is more conducive to pulmonary drug delivery and disease control than traditional systemic drug delivery, significantly circumventing drug loss, off-target effects, systemic and organ toxicity, etc., and is widely regarded as the preferred regimen for pulmonary drug delivery. However, very few lung diseases are currently treated with the preferred inhaled formulations, such as asthma, chronic obstructive pulmonary disease and pulmonary hypertension. And there is a lack of appropriate inhaled formulations for other critical lung diseases, such as lung cancer and pulmonary fibrosis, due to the fact that the physicochemical properties of the drugs and their pharmacokinetic profiles do not match the physiology of the lung, and conventional inhalation devices are unable to deliver them to the specific parts of the lung. Phytochemicals of natural origin, due to their wide availability and clear safety profile, hold great promise for the preparation of inhalable formulations to improve the current dilemma in the treatment of lung diseases. In particular, the preparation of inhalable formulations based on

According to the hypothesis that strange quark matter may be the true ground state of matter at extremely high densities, strange quark stars should be stable and could exist in the Universe. It is possible that pulsars may actually be strange stars, but not neutron stars. Here we present a short review on recent progresses in the field of strange quark stars. Three popular phenomenological models widely used to describe strange quark matter are introduced, with special attention being paid on the corresponding equation of state in each model. Combining the equation of state with the Tolman-Oppenheimer-Volkov equations, the inner structure and mass-radius relation can be obtained for the whole sequence of strange stars. Strong gravitational wave emissions may be generated by strange stars through various mechanisms, which may help identify strange stars via observations. Especially, close-in strange quark planets with respect to their hosts may provide a unique test for the existence of strange quark objects. Fierce electromagnetic bursts could also be generated by strange stars. The energy may come from the phase transition of neutron stars to strange stars, or from the merger of binary strange stars. The collapse of the strange star crust can also release a huge amount of energy. It is shown that strange quark stars may be involved in short gamma-ray bursts and fast radio bursts.

Red dwarfs have been suggested to be among the possible astrophysical species accelerating particles and emitting TeV $\gamma$-rays. As an effort to search for the GeV $\gamma$-ray counterparts of the suggested TeV emission from eight red dwarfs, we analyse the 0.2--500 GeV $\gamma$-ray emission of the regions covering them exploiting the $\sim$13.6 yr Pass 8 data of the Fermi Large Area Telescope. A GeV $\gamma$-ray emission excess with significance of 3.8$\sigma$ is detected in the direction of the red dwarf V962 Tau. This emission contains V962 Tau in 1$\sigma$ error radius and is independent of the catalog source. However, the stellar flare scenario can hardly explain the total energy and lightcurve derived from the $\gamma$-ray emission in view of the spectral analysis. We also analyse the lightcurves in the positions of the eight red dwarfs and no time bin with significance $>$5$\sigma$ is found. Therefore, no significant emission from the red dwarfs could be concluded to be detected by Fermi-LAT., Comment: 11 pages, 6 figures

Radio frequency interference (RFI) have been an enduring concern in radio astronomy, particularly for the observations of pulsars which require high timing precision and data sensitivity. In most works of the literature, RFI mitigation has been formulated as a detection task that consists of localizing possible RFI in dynamic spectra. This strategy inevitably leads to a potential loss of information since parts of the signal identified as possibly RFI-corrupted are generally not considered in the subsequent data processing pipeline. Conversely, this work proposes to tackle RFI mitigation as a joint detection and restoration that allows parts of the dynamic spectrum affected by RFI to be not only identified but also recovered. The proposed supervised method relies on a deep convolutional network whose architecture inherits the performance reached by a recent yet popular image-denoising network. To train this network, a whole simulation framework is built to generate large data sets according to physics-inspired and statistical models of the pulsar signals and of the RFI. The relevance of the proposed approach is quantitatively assessed by conducting extensive experiments. In particular, the results show that the restored dynamic spectra are sufficiently reliable to estimate pulsar times-of-arrivals with an accuracy close to the one that would be obtained from RFI-free signals.

According to the strange quark matter hypothesis, strange planets may exist, which are planetary mass objects composed of almost equal numbers of up, down and strange quarks. A strange planet can revolve around its host strange star in a very close-in orbit. When it finally merges with the host, strong gravitational wave emissions will be generated. Here the gravitational waveforms are derived for the merging process, taking into account the effects of the strange star's magnetic field on the dynamics. Effects of the inclination angle are also considered. Templates of the gravitational waveforms are derived. It is found that the magnetic interactions significantly speed up the merging process. Coalescence events of such strange planetary systems occurring in our Galaxy as well as in local galaxies can be effectively detected by current and future gravitational experiments, which may hopefully provide a new method to test the strange quark matter hypothesis and probe the magnetic field of compact stars.

Background: Academic emergency medicine (EM) communities have viewed anonymous online communities (AOC) such as Reddit or specialty-specific “applicant spreadsheets” as poor advising resources. Despite this, robust EM AOCs exist, with large user bases and heavy readership. Insights about applicants’ authentic experiences can be critical for applicants and program leadership decision-making. To date, there are no EM studies to qualitatively assess EM AOC narratives during the application cycle. Our goal was to perform a qualitative analysis of students’ EM program experiences through a publicly available AOC. Methods: This was a qualitative analysis of a publicly available, time-stamped, user-locked AOC dataset: “Official 2020–2021 Emergency Medicine Applicant Spreadsheet.” We extracted and then deidentified all data from selected sub-sheets entitled “Virtual Interview Impressions” and “Rotation Impressions.” Four investigators used constant comparative method to analyze the data inductively, and they subsequently met to generate common themes discussed by students. Preliminary thematic analysis was conducted on a random sample of 37/183 (20%) independent narratives to create the initial codebook. This was used and updated iteratively to analyze the entire narrative set consisting of 841 discrete statements. Finally, two unique codes were created to distinguish whether the identified subthemes, or program attributes, were likely “modifiable” or “non-modifiable.” Results: We identified six major themes: living and working conditions; interpersonal relationships; learning experiences, postgraduate readiness, and online/virtual supplements. Common sub-themes included patient population (13%); resident personality (7%); program leadership personality (7%); relationship with faculty/ leadership (6%); geography (4%); practice setting (4%); program reputation (4%), and postgraduate year-3 experiences (4%). Modifiable sub-themes outnumbered non-modifiable sub-themes, 60.7% t

Recent advancements in Large Language Models (LLMs) have attracted considerable interest among researchers to leverage these models to enhance Recommender Systems (RSs). Existing work predominantly utilizes LLMs to generate knowledge-rich texts or utilizes LLM-derived embeddings as features to improve RSs. Although the extensive world knowledge embedded in LLMs generally benefits RSs, the application can only take limited number of users and items as inputs, without adequately exploiting collaborative filtering information. Considering its crucial role in RSs, one key challenge in enhancing RSs with LLMs lies in providing better collaborative filtering information through LLMs. In this paper, drawing inspiration from the in-context learning and chain of thought reasoning in LLMs, we propose the Large Language Models enhanced Collaborative Filtering (LLM-CF) framework, which distils the world knowledge and reasoning capabilities of LLMs into collaborative filtering. We also explored a concise and efficient instruction-tuning method, which improves the recommendation capabilities of LLMs while preserving their general functionalities (e.g., not decreasing on the LLM benchmark). Comprehensive experiments on three real-world datasets demonstrate that LLM-CF significantly enhances several backbone recommendation models and consistently outperforms competitive baselines, showcasing its effectiveness in distilling the world knowledge and reasoning capabilities of LLM into collaborative filtering., Comment: 11 pages

Traditional imitation learning focuses on modeling the behavioral mechanisms of experts, which requires a large amount of interaction history generated by some fixed expert. However, in many streaming applications, such as streaming recommender systems, online decision-makers typically engage in online learning during the decision-making process, meaning that the interaction history generated by online decision-makers includes their behavioral evolution from novice expert to experienced expert. This poses a new challenge for existing imitation learning approaches that can only utilize data from experienced experts. To address this issue, this paper proposes an inverse batched contextual bandit (IBCB) framework that can efficiently perform estimations of environment reward parameters and learned policy based on the expert's behavioral evolution history. Specifically, IBCB formulates the inverse problem into a simple quadratic programming problem by utilizing the behavioral evolution history of the batched contextual bandit with inaccessible rewards. We demonstrate that IBCB is a unified framework for both deterministic and randomized bandit policies. The experimental results indicate that IBCB outperforms several existing imitation learning algorithms on synthetic and real-world data and significantly reduces running time. Additionally, empirical analyses reveal that IBCB exhibits better out-of-distribution generalization and is highly effective in learning the bandit policy from the interaction history of novice experts.

The deepfake threats to society and cybersecurity have provoked significant public apprehension, driving intensified efforts within the realm of deepfake video detection. Current video-level methods are mostly based on {3D CNNs} resulting in high computational demands, although have achieved good performance. This paper introduces an elegantly simple yet effective strategy named Thumbnail Layout (TALL), which transforms a video clip into a pre-defined layout to realize the preservation of spatial and temporal dependencies. This transformation process involves sequentially masking frames at the same positions within each frame. These frames are then resized into sub-frames and reorganized into the predetermined layout, forming thumbnails. TALL is model-agnostic and has remarkable simplicity, necessitating only minimal code modifications. Furthermore, we introduce a graph reasoning block (GRB) and semantic consistency (SC) loss to strengthen TALL, culminating in TALL++. GRB enhances interactions between different semantic regions to capture semantic-level inconsistency clues. The semantic consistency loss imposes consistency constraints on semantic features to improve model generalization ability. Extensive experiments on intra-dataset, cross-dataset, diffusion-generated image detection, and deepfake generation method recognition show that TALL++ achieves results surpassing or comparable to the state-of-the-art methods, demonstrating the effectiveness of our approaches for various deepfake detection problems. The code is available at https://github.com/rainy-xu/TALL4Deepfake., Comment: Accepted by IJCV

A heterogeneous micro aerial vehicles (MAV) swarm consists of resource-intensive but expensive advanced MAVs (AMAVs) and resource-limited but cost-effective basic MAVs (BMAVs), offering opportunities in diverse fields. Accurate and real-time localization is crucial for MAV swarms, but current practices lack a low-cost, high-precision, and real-time solution, especially for lightweight BMAVs. We find an opportunity to accomplish the task by transforming AMAVs into mobile localization infrastructures for BMAVs. However, turning this insight into a practical system is non-trivial due to challenges in location estimation with BMAVs' unknown and diverse localization errors and resource allocation of AMAVs given coupled influential factors. This study proposes TransformLoc, a new framework that transforms AMAVs into mobile localization infrastructures, specifically designed for low-cost and resource-constrained BMAVs. We first design an error-aware joint location estimation model to perform intermittent joint location estimation for BMAVs and then design a proximity-driven adaptive grouping-scheduling strategy to allocate resources of AMAVs dynamically. TransformLoc achieves a collaborative, adaptive, and cost-effective localization system suitable for large-scale heterogeneous MAV swarms. We implement TransformLoc on industrial drones and validate its performance. Results show that TransformLoc outperforms baselines including SOTA up to 68\% in localization performance, motivating up to 60\% navigation success rate improvement., Comment: 10 pages, accepted by IEEE INFOCOM 2024

B cell acute lymphoblastic leukemia (ALL) is characterized by the highly heterogeneity of pathogenic genetic background, and there are still approximately 30–40% of patients without clear molecular markers. To identify the dysregulated genes in B cell ALL, we screened 30 newly diagnosed B cell ALL patients and 10 donors by gene expression profiling chip. We found that ECM1 transcription level was abnormally elevated in newly diagnosed B cell ALL and further verified in another 267 cases compared with donors (median, 124.57% vs. 7.14%, P < 0.001). ROC analysis showed that the area under the curve of ECM1 transcription level at diagnosis was 0.89 (P < 0.001). Patients with BCR::ABL1 and IKZF1 deletion show highest transcription level (210.78%) compared with KMT2A rearrangement (39.48%) and TCF3::PBX1 rearrangement ones (30.02%) (all P < 0.05). Also, the transcription level of ECM1 was highly correlated with the clinical course, as 20 consecutive follow-up cases indicated. The 5-year OS of patients (non-KMT2A and non-TCF3::PBX1 rearrangement) with high ECM1 transcription level was significantly worse than the lower ones (18.7% vs. 72.9%, P < 0.001) and high ECM1 transcription level was an independent risk factor for OS (HR = 5.77 [1.75–19.06], P = 0.004). After considering transplantation, high ECM1 transcription level was not an independent risk factor, although OS was still poor (low vs. high, 71.1% vs. 56.8%, P = 0.038). Our findings suggested that ECM1 may be a potential molecular marker for diagnosis, minimal residual disease (MRD) monitoring, and prognosis prediction of B cell ALL. Trial registration Trial Registration Registered in the Beijing Municipal Health Bureau Registration N 2007–1007 and in the Chinese Clinical Trial Registry [ChiCTR-OCH-10000940 and ChiCTR-OPC-14005546]; http://www.chictr.org.cn. © The Author(s) 2024.

The successful application of gradient boosting regression (GBR) in machine learning to forecast surface area, pore volume, and yield in biomass-derived activated carbon (AC) production underscores its potential for enhancing manufacturing processes. The GBR model, collecting 17 independent variables for two-step activation (2-SA) and 14 for one-step activation (1-SA), demonstrates effectiveness across three datasets—1-SA, 2-SA, and a combined dataset. Notably, in 1-SA, the GBR model yields R2 values of 0.76, 0.90, and 0.83 for TPV, yield, and SSA respectively, and records R2 of 0.90 and 0.91 for yield in 2-SA and combined datasets. The model highlights the significance of the soaking procedure alongside activation temperature in shaping AC properties with 1-SA or 2-SA, illustrating machine learning's potential in optimizing AC production processes.

In this work, we present RadCloud, a novel real time framework for directly obtaining higher-resolution lidar-like 2D point clouds from low-resolution radar frames on resource-constrained platforms commonly used in unmanned aerial and ground vehicles (UAVs and UGVs, respectively); such point clouds can then be used for accurate environmental mapping, navigating unknown environments, and other robotics tasks. While high-resolution sensing using radar data has been previously reported, existing methods cannot be used on most UAVs, which have limited computational power and energy; thus, existing demonstrations focus on offline radar processing. RadCloud overcomes these challenges by using a radar configuration with 1/4th of the range resolution and employing a deep learning model with 2.25x fewer parameters. Additionally, RadCloud utilizes a novel chirp-based approach that makes obtained point clouds resilient to rapid movements (e.g., aggressive turns or spins), which commonly occur during UAV flights. In real-world experiments, we demonstrate the accuracy and applicability of RadCloud on commercially available UAVs and UGVs, with off-the-shelf radar platforms on-board., Comment: $\copyright$ 2024 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works

Despite extensive pre-training and fine-tuning in moral alignment to prevent generating harmful information at user request, large language models (LLMs) remain vulnerable to jailbreak attacks. In this paper, we propose AutoDefense, a response-filtering based multi-agent defense framework that filters harmful responses from LLMs. This framework assigns different roles to LLM agents and employs them to complete the defense task collaboratively. The division in tasks enhances the overall instruction-following of LLMs and enables the integration of other defense components as tools. AutoDefense can adapt to various sizes and kinds of open-source LLMs that serve as agents. Through conducting extensive experiments on a large scale of harmful and safe prompts, we validate the effectiveness of the proposed AutoDefense in improving the robustness against jailbreak attacks, while maintaining the performance at normal user request. Our code and data are publicly available at https://github.com/XHMY/AutoDefense.

Building on the success of text-to-image diffusion models (DPMs), image editing is an important application to enable human interaction with AI-generated content. Among various editing methods, editing within the prompt space gains more attention due to its capacity and simplicity of controlling semantics. However, since diffusion models are commonly pretrained on descriptive text captions, direct editing of words in text prompts usually leads to completely different generated images, violating the requirements for image editing. On the other hand, existing editing methods usually consider introducing spatial masks to preserve the identity of unedited regions, which are usually ignored by DPMs and therefore lead to inharmonic editing results. Targeting these two challenges, in this work, we propose to disentangle the comprehensive image-prompt interaction into several item-prompt interactions, with each item linked to a special learned prompt. The resulting framework, named D-Edit, is based on pretrained diffusion models with cross-attention layers disentangled and adopts a two-step optimization to build item-prompt associations. Versatile image editing can then be applied to specific items by manipulating the corresponding prompts. We demonstrate state-of-the-art results in four types of editing operations including image-based, text-based, mask-based editing, and item removal, covering most types of editing applications, all within a single unified framework. Notably, D-Edit is the first framework that can (1) achieve item editing through mask editing and (2) combine image and text-based editing. We demonstrate the quality and versatility of the editing results for a diverse collection of images through both qualitative and quantitative evaluations.

Federated learning (FL) is a novel distributed machine learning paradigm that enables participants to collaboratively train a centralized model with privacy preservation by eliminating the requirement of data sharing. In practice, FL often involves multiple participants and requires the third party to aggregate global information to guide the update of the target participant. Therefore, many FL methods do not work well due to the training and test data of each participant may not be sampled from the same feature space and the same underlying distribution. Meanwhile, the differences in their local devices (system heterogeneity), the continuous influx of online data (incremental data), and labeled data scarcity may further influence the performance of these methods. To solve this problem, federated transfer learning (FTL), which integrates transfer learning (TL) into FL, has attracted the attention of numerous researchers. However, since FL enables a continuous share of knowledge among participants with each communication round while not allowing local data to be accessed by other participants, FTL faces many unique challenges that are not present in TL. In this survey, we focus on categorizing and reviewing the current progress on federated transfer learning, and outlining corresponding solutions and applications. Furthermore, the common setting of FTL scenarios, available datasets, and significant related research are summarized in this survey.

Large language models (LLMs) are now increasingly utilized for role-playing tasks, especially in impersonating domain-specific experts, primarily through role-playing prompts. When interacting in real-world scenarios, the decision-making abilities of a role significantly shape its behavioral patterns. In this paper, we concentrate on evaluating the decision-making abilities of LLMs post role-playing thereby validating the efficacy of role-playing. Our goal is to provide metrics and guidance for enhancing the decision-making abilities of LLMs in role-playing tasks. Specifically, we first use LLMs to generate virtual role descriptions corresponding to the 16 personality types of Myers-Briggs Type Indicator (abbreviated as MBTI) representing a segmentation of the population. Then we design specific quantitative operations to evaluate the decision-making abilities of LLMs post role-playing from four aspects: adaptability, exploration$\&$exploitation trade-off ability, reasoning ability, and safety. Finally, we analyze the association between the performance of decision-making and the corresponding MBTI types through GPT-4. Extensive experiments demonstrate stable differences in the four aspects of decision-making abilities across distinct roles, signifying a robust correlation between decision-making abilities and the roles emulated by LLMs. These results underscore that LLMs can effectively impersonate varied roles while embodying their genuine sociological characteristics.

LLMs have transformed NLP and shown promise in various fields, yet their potential in finance is underexplored due to a lack of thorough evaluations and the complexity of financial tasks. This along with the rapid development of LLMs, highlights the urgent need for a systematic financial evaluation benchmark for LLMs. In this paper, we introduce FinBen, the first comprehensive open-sourced evaluation benchmark, specifically designed to thoroughly assess the capabilities of LLMs in the financial domain. FinBen encompasses 35 datasets across 23 financial tasks, organized into three spectrums of difficulty inspired by the Cattell-Horn-Carroll theory, to evaluate LLMs' cognitive abilities in inductive reasoning, associative memory, quantitative reasoning, crystallized intelligence, and more. Our evaluation of 15 representative LLMs, including GPT-4, ChatGPT, and the latest Gemini, reveals insights into their strengths and limitations within the financial domain. The findings indicate that GPT-4 leads in quantification, extraction, numerical reasoning, and stock trading, while Gemini shines in generation and forecasting; however, both struggle with complex extraction and forecasting, showing a clear need for targeted enhancements. Instruction tuning boosts simple task performance but falls short in improving complex reasoning and forecasting abilities. FinBen seeks to continuously evaluate LLMs in finance, fostering AI development with regular updates of tasks and models., Comment: 19 pages, 10 figures

In pursuit of fairness and balanced development, recommender systems (RS) often prioritize group fairness, ensuring that specific groups maintain a minimum level of exposure over a given period. For example, RS platforms aim to ensure adequate exposure for new providers or specific categories of items according to their needs. Modern industry RS usually adopts a two-stage pipeline: stage-1 (retrieval stage) retrieves hundreds of candidates from millions of items distributed across various servers, and stage-2 (ranking stage) focuses on presenting a small-size but accurate selection from items chosen in stage-1. Existing efforts for ensuring amortized group exposures focus on stage-2, however, stage-1 is also critical for the task. Without a high-quality set of candidates, the stage-2 ranker cannot ensure the required exposure of groups. Previous fairness-aware works designed for stage-2 typically require accessing and traversing all items. In stage-1, however, millions of items are distributively stored in servers, making it infeasible to traverse all of them. How to ensure group exposures in the distributed retrieval process is a challenging question. To address this issue, we introduce a model named FairSync, which transforms the problem into a constrained distributed optimization problem. Specifically, FairSync resolves the issue by moving it to the dual space, where a central node aggregates historical fairness data into a vector and distributes it to all servers. To trade off the efficiency and accuracy, the gradient descent technique is used to periodically update the parameter of the dual vector. The experiment results on two public recommender retrieval datasets showcased that FairSync outperformed all the baselines, achieving the desired minimum level of exposures while maintaining a high level of retrieval accuracy., Comment: Accepted in WWW'24

Despite Spanish's pivotal role in the global finance industry, a pronounced gap exists in Spanish financial natural language processing (NLP) and application studies compared to English, especially in the era of large language models (LLMs). To bridge this gap, we unveil Tois\'on de Oro, the first bilingual framework that establishes instruction datasets, finetuned LLMs, and evaluation benchmark for financial LLMs in Spanish joint with English. We construct a rigorously curated bilingual instruction dataset including over 144K Spanish and English samples from 15 datasets covering 7 tasks. Harnessing this, we introduce FinMA-ES, an LLM designed for bilingual financial applications. We evaluate our model and existing LLMs using FLARE-ES, the first comprehensive bilingual evaluation benchmark with 21 datasets covering 9 tasks. The FLARE-ES benchmark results reveal a significant multilingual performance gap and bias in existing LLMs. FinMA-ES models surpass SOTA LLMs such as GPT-4 in Spanish financial tasks, due to strategic instruction tuning and leveraging data from diverse linguistic resources, highlighting the positive impact of cross-linguistic transfer. All our datasets, models, and benchmarks have been released., Comment: 10 pages, 2 figures

Crowd simulation holds crucial applications in various domains, such as urban planning, architectural design, and traffic arrangement. In recent years, physics-informed machine learning methods have achieved state-of-the-art performance in crowd simulation but fail to model the heterogeneity and multi-modality of human movement comprehensively. In this paper, we propose a social physics-informed diffusion model named SPDiff to mitigate the above gap. SPDiff takes both the interactive and historical information of crowds in the current timeframe to reverse the diffusion process, thereby generating the distribution of pedestrian movement in the subsequent timeframe. Inspired by the well-known social physics model, i.e., Social Force, regarding crowd dynamics, we design a crowd interaction module to guide the denoising process and further enhance this module with the equivariant properties of crowd interactions. To mitigate error accumulation in long-term simulations, we propose a multi-frame rollout training algorithm for diffusion modeling. Experiments conducted on two real-world datasets demonstrate the superior performance of SPDiff in terms of macroscopic and microscopic evaluation metrics. Code and appendix are available at https://github.com/tsinghua-fib-lab/SPDiff.

To foster an immersive and natural human-robot interaction, the implementation of tactile perception and feedback becomes imperative, effectively bridging the conventional sensory gap. In this paper, we propose a dual-modal electronic skin (e-skin) that integrates magnetic tactile sensing and vibration feedback for enhanced human-robot interaction. The dual-modal tactile e-skin offers multi-functional tactile sensing and programmable haptic feedback, underpinned by a layered structure comprised of flexible magnetic films, soft silicone, a Hall sensor and actuator array, and a microcontroller unit. The e-skin captures the magnetic field changes caused by subtle deformations through Hall sensors, employing deep learning for accurate tactile perception. Simultaneously, the actuator array generates mechanical vibrations to facilitate haptic feedback, delivering diverse mechanical stimuli. Notably, the dual-modal e-skin is capable of transmitting tactile information bidirectionally, enabling object recognition and fine-weighing operations. This bidirectional tactile interaction framework will enhance the immersion and efficiency of interactions between humans and robots., Comment: 7 pages, 8 figures. Submitted to 2024 IEEE International Conference on Robotics and Automation (ICRA), Japan, Yokohama

Reinforcement learning (RL) with sparse and deceptive rewards is challenging because non-zero rewards are rarely obtained. Hence, the gradient calculated by the agent can be stochastic and without valid information. Recent studies that utilize memory buffers of previous experiences can lead to a more efficient learning process. However, existing methods often require these experiences to be successful and may overly exploit them, which can cause the agent to adopt suboptimal behaviors. This paper develops an approach that uses diverse past trajectories for faster and more efficient online RL, even if these trajectories are suboptimal or not highly rewarded. The proposed algorithm combines a policy improvement step with an additional exploration step using offline demonstration data. The main contribution of this paper is that by regarding diverse past trajectories as guidance, instead of imitating them, our method directs its policy to follow and expand past trajectories while still being able to learn without rewards and approach optimality. Furthermore, a novel diversity measurement is introduced to maintain the team's diversity and regulate exploration. The proposed algorithm is evaluated on discrete and continuous control tasks with sparse and deceptive rewards. Compared with the existing RL methods, the experimental results indicate that our proposed algorithm is significantly better than the baseline methods regarding diverse exploration and avoiding local optima., Comment: 23 pages, 19 figures

To tackle the issues of catastrophic forgetting and overfitting in few-shot class-incremental learning (FSCIL), previous work has primarily concentrated on preserving the memory of old knowledge during the incremental phase. The role of pre-trained model in shaping the effectiveness of incremental learning is frequently underestimated in these studies. Therefore, to enhance the generalization ability of the pre-trained model, we propose Learning with Prior Knowledge (LwPK) by introducing nearly free prior knowledge from a few unlabeled data of subsequent incremental classes. We cluster unlabeled incremental class samples to produce pseudo-labels, then jointly train these with labeled base class samples, effectively allocating embedding space for both old and new class data. Experimental results indicate that LwPK effectively enhances the model resilience against catastrophic forgetting, with theoretical analysis based on empirical risk minimization and class distance measurement corroborating its operational principles. The source code of LwPK is publicly available at: \url{https://github.com/StevenJ308/LwPK}.

Most vision-based tactile sensors use elastomer deformation to infer tactile information, which can not sense some modalities, like temperature. As an important part of human tactile perception, temperature sensing can help robots better interact with the environment. In this work, we propose a novel multimodal vision-based tactile sensor, SATac, which can simultaneously perceive information of temperature, pressure, and shear. SATac utilizes thermoluminescence of strontium aluminate (SA) to sense a wide range of temperatures with exceptional resolution. Additionally, the pressure and shear can also be perceived by analyzing Voronoi diagram. A series of experiments are conducted to verify the performance of our proposed sensor. We also discuss the possible application scenarios and demonstrate how SATac could benefit robot perception capabilities.