Search

Li-Wei Zhang,1– 3,* Man-Qing Luo,4,* Ji-Lang Zeng,1– 3,* Zhe-Bin You,2,3,5 Li-Chuan Chen,1– 3 Jun-Han Chen,1– 3 Kai-Yang Lin,1– 3 Yan-Song Guo1– 3 1Department of Cardiology, Shengli Clinical Medical College of Fujian Medical University, Fujian Provincial Hospital, Fuzhou, People’s Republic of China; 2Fujian Provincial Key Laboratory of Cardiovascular Disease, Fujian Provincial Center for Geriatrics, Fujian Provincial Clinical Research Center for Severe Acute Cardiovascular Diseases, Fuzhou, People’s Republic of China; 3Fujian Heart Failure Center Alliance, Fuzhou, People’s Republic of China; 4Department of Cardiology, Fuwai Hospital, National Clinical Research Center for Cardiovascular Diseases, National Center for Cardiovascular Diseases, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, People’s Republic of China; 5Department of Geriatric Medicine, Fujian Key Laboratory of Geriatrics, Fujian Provincial Hospital, Fujian Provincial Center for Geriatrics, Fujian Medical University, Fuzhou, Fujian, 350001, People’s Republic of China*These authors contributed equally to this workCorrespondence: Kai-Yang Lin; Yan-Song Guo, Tel +86-13559355708 ; +86-059188217410, Email lky7411@sina.com; ysguo1234@126.comPurpose: The estimated glomerular filtration rate (eGFR) based on creatinine is crucial for the risk assessment of contrast-associated acute kidney injury (CA-AKI). In recent, the difference between cystatin C-based eGFR (eGFRcys) and creatinine-based eGFR (eGFRcr) has been widely documented. We aimed to explore whether intraindividual differences between eGFRcys and eGFRcr had potential value for CA-AKI risk assessment in patients undergoing elective percutaneous coronary intervention (PCI).Patients and Methods: From January 2012 to December 2018, we retrospectively observed 5049 patients receiving elective PCI. To determine eGFR, serum creatinine and cystatin C levels were measured. CA-AK

STEAM education system is largely being incorporated within different countries to develop creativity and problem-solving skills among the students. This study has also focused on sustaining regional equality and interdisciplinary integration of arts and sciences via creativity research in STEAM education systems. For this study, the qualitative data was collected from 12 science teachers working in different Chinese high schools. Thematic analysis was used for analyzing the collected interview transcripts. Four important themes were formulated which include (a) STEAM Education System, (b) Creativity in STEAM Education System, (c) STEAM Education System Issues and (d) Recommendations for STEAM Education System. The results obtained from this study showed that STEAM education system helps in encouraging student engagement, following interdisciplinary integration of arts and sciences. However, regional inequality in China prevents effective accessibility and quality of STEAM education in rural regions of the country, emphasizing the development and implementation of effective educational and regional policies. This study has also provided different practical and theoretical implications within the context of STEAM education systems., El sistema educativo STEAM se está incorporando en diferentes países principalmente para desarrollar la creatividad y la capacidad de resolución de problemas entre los estudiantes. Este estudio se ha centrado en mantener la igualdad regional y la integración interdisciplinar de las artes y las ciencias a través de la investigación de la creatividad en los sistemas educativos STEAM. Para este estudio, se recogieron datos cualitativos de 12 profesores de ciencias que trabajan en diferentes bachilleratos chinos. Se utilizó el análisis temático para analizar las transcripciones de las entrevistas recogidas. Se formularon cuatro temas importantes que incluyen (a) Sistema educativo STEAM, (b) Creatividad en el sistema educativo STEAM, (c) Problemas del sistema educativo STEAM y (d) Recomendaciones para el sistema educativo STEAM. Los resultados obtenidos en este estudio muestran que el sistema educativo STEAM ayuda a fomentar el compromiso de los estudiantes, tras la integración interdisciplinar de las artes y las ciencias. Sin embargo, la desigualdad regional en China impide la accesibilidad efectiva y la calidad de la educación STEAM en las regiones rurales del país, haciendo hincapié en el desarrollo y la aplicación de políticas educativas y regionales eficaces. Este estudio también ha aportado diferentes implicaciones prácticas y teóricas en el contexto de los sistemas educativos STEAM.

Over the past few years, Text-to-Image (T2I) generation approaches based on diffusion models have gained significant attention. However, vanilla diffusion models often suffer from spelling inaccuracies in the text displayed within the generated images. The capability to generate visual text is crucial, offering both academic interest and a wide range of practical applications. To produce accurate visual text images, state-of-the-art techniques adopt a glyph-controlled image generation approach, consisting of a text layout generator followed by an image generator that is conditioned on the generated text layout. Nevertheless, our study reveals that these models still face three primary challenges, prompting us to develop a testbed to facilitate future research. We introduce a benchmark, LenCom-Eval, specifically designed for testing models' capability in generating images with Lengthy and Complex visual text. Subsequently, we introduce a training-free framework to enhance the two-stage generation approaches. We examine the effectiveness of our approach on both LenCom-Eval and MARIO-Eval benchmarks and demonstrate notable improvements across a range of evaluation metrics, including CLIPScore, OCR precision, recall, F1 score, accuracy, and edit distance scores. For instance, our proposed framework improves the backbone model, TextDiffuser, by more than 23\% and 13.5\% in terms of OCR word F1 on LenCom-Eval and MARIO-Eval, respectively. Our work makes a unique contribution to the field by focusing on generating images with long and rare text sequences, a niche previously unexplored by existing literature

Autoregressive Visual Language Models (VLMs) showcase impressive few-shot learning capabilities in a multimodal context. Recently, multimodal instruction tuning has been proposed to further enhance instruction-following abilities. However, we uncover the potential threat posed by backdoor attacks on autoregressive VLMs during instruction tuning. Adversaries can implant a backdoor by injecting poisoned samples with triggers embedded in instructions or images, enabling malicious manipulation of the victim model's predictions with predefined triggers. Nevertheless, the frozen visual encoder in autoregressive VLMs imposes constraints on the learning of conventional image triggers. Additionally, adversaries may encounter restrictions in accessing the parameters and architectures of the victim model. To address these challenges, we propose a multimodal instruction backdoor attack, namely VL-Trojan. Our approach facilitates image trigger learning through an isolating and clustering strategy and enhance black-box-attack efficacy via an iterative character-level text trigger generation method. Our attack successfully induces target outputs during inference, significantly surpassing baselines (+62.52\%) in ASR. Moreover, it demonstrates robustness across various model scales and few-shot in-context reasoning scenarios.

Language models, especially pre-trained large language models, have showcased remarkable abilities as few-shot in-context learners (ICL), adept at adapting to new tasks with just a few demonstrations in the input context. However, the model's ability to perform ICL is sensitive to the choice of the few-shot demonstrations. Instead of using a fixed set of demonstrations, one recent development is to retrieve demonstrations tailored to each input query. The implementation of demonstration retrieval is relatively straightforward, leveraging existing databases and retrieval systems. This not only improves the efficiency and scalability of the learning process but also has been shown to reduce biases inherent in manual example selection. In light of the encouraging results and growing research in ICL with retrieved demonstrations, we conduct an extensive review of studies in this area. In this survey, we discuss and compare different design choices for retrieval models, retrieval training procedures, and inference algorithms.

The integration of Large Language Models (LLMs) into the healthcare domain has the potential to significantly enhance patient care and support through the development of empathetic, patient-facing chatbots. This study investigates an intriguing question Can ChatGPT respond with a greater degree of empathy than those typically offered by physicians? To answer this question, we collect a de-identified dataset of patient messages and physician responses from Mayo Clinic and generate alternative replies using ChatGPT. Our analyses incorporate novel empathy ranking evaluation (EMRank) involving both automated metrics and human assessments to gauge the empathy level of responses. Our findings indicate that LLM-powered chatbots have the potential to surpass human physicians in delivering empathetic communication, suggesting a promising avenue for enhancing patient care and reducing professional burnout. The study not only highlights the importance of empathy in patient interactions but also proposes a set of effective automatic empathy ranking metrics, paving the way for the broader adoption of LLMs in healthcare.

Recently developed large language models (LLMs) have been shown to perform remarkably well on a wide range of language understanding tasks. But, can they really "reason" over the natural language? This question has been receiving significant research attention and many reasoning skills such as commonsense, numerical, and qualitative have been studied. However, the crucial skill pertaining to 'logical reasoning' has remained underexplored. Existing work investigating this reasoning ability of LLMs has focused only on a couple of inference rules (such as modus ponens and modus tollens) of propositional and first-order logic. Addressing the above limitation, we comprehensively evaluate the logical reasoning ability of LLMs on 25 different reasoning patterns spanning over propositional, first-order, and non-monotonic logics. To enable systematic evaluation, we introduce LogicBench, a natural language question-answering dataset focusing on the use of a single inference rule. We conduct detailed analysis with a range of LLMs such as GPT-4, ChatGPT, Gemini, Llama-2, and Mistral using chain-of-thought prompting. Experimental results show that existing LLMs do not fare well on LogicBench; especially, they struggle with instances involving complex reasoning and negations. Furthermore, they sometimes overlook contextual information necessary for reasoning to arrive at the correct conclusion. We believe that our work and findings facilitate future research for evaluating and enhancing the logical reasoning ability of LLMs. Data and code are available at https://github.com/Mihir3009/LogicBench., Comment: Accepted at ACL(Main) 2024 | First version available @ https://openreview.net/forum?id=7NR2ZVzZxx

Sea spray aerosol (SSA) ejected through bursting bubbles at the ocean surface is a complex mixture of salts and organic species. Submicrometer SSA particles have long atmospheric lifetimes and play a critical role in the climate system. Composition impacts their ability to form marine clouds, yet their cloud-forming potential is difficult to study due to their small size. Here, we use large-scale molecular dynamics (MD) simulations as a "computational microscope" to provide never-before-seen views of 40 nm model aerosol particles and their molecular morphologies. We investigate how increasing chemical complexity impacts the distribution of organic material throughout individual particles for a range of organic constituents with varying chemical properties. Our simulations show that common organic marine surfactants readily partition between both the surface and interior of the aerosol, indicating that nascent SSA may be more heterogeneous than traditional morphological models suggest. We support our computational observations of SSA surface heterogeneity with Brewster angle microscopy on model interfaces. These observations indicate that increased chemical complexity in submicrometer SSA leads to a reduced surface coverage by marine organics, which may facilitate water uptake in the atmosphere. Our work thus establishes large-scale MD simulations as a novel technique for interrogating aerosols at the single-particle level.

Grand Chinese infrastructure projects are the key to Myanmar’s plan to bridge the major gaps in the transportation and energy infrastructure essential for its economic development. However, the design and execution of Chinese infrastructure projects have been criticized over their unclear benefits for Myanmar, high project costs, and disruptive impacts on local ecology and culture. This chapter reviews the mixed local receptions to Chinese investment, drawing on fieldwork in Myanmar, and suggests the potential for leveraging Hong Kong’s managerial and professional experience in enhancing responsible investment in the Belt and Road Initiative.

Aim: A population pharmacokinetic (PPK) model was developed to characterize pharmacokinetics (PK) of subcutaneous or intravenous daratumumab administration in a new indication (i.e., combination with pomalidomide and dexamethasone [D-Pd] in patients with relapsed or refractory multiple myeloma [RRMM]). Analyses were conducted to explore exposure–response (E-R) relationships for efficacy and select treatment-emergent adverse events (TEAEs). Methods: The PPK analysis included pooled data from the D-Pd cohorts of the phase 3 APOLLO and phase 1b EQUULEUS studies. Covariates were evaluated in the PPK model. Model-predicted exposures to daratumumab were compared between covariate subgroups of interest and used to investigate relationships between daratumumab exposure and efficacy and safety in APOLLO. Results: The PPK analysis included 1146 daratumumab PK samples from 239 patients (APOLLO, n = 140; EQUULEUS, n = 99). Observed concentration–time data of daratumumab were well described by a two-compartment PPK model with first-order absorption and parallel linear and nonlinear elimination pathways. Treatment with D-Pd provided similar daratumumab PK characteristics versus historical daratumumab monotherapy. The E-R dataset contained data from 290 APOLLO patients (D-Pd, n = 140; Pd, n = 150). The PK–efficacy relationship of daratumumab supported improved progression-free survival for patients in the D-Pd group vs. the Pd group. Additionally, TEAEs did not increase with increasing PK exposure in the D-Pd group. Conclusions: The PPK and E-R analyses support the daratumumab subcutaneous 1800 mg dosing regimen in combination with Pd for treatment of patients with RRMM. No dose adjustment is recommended in this indication for any of the investigated factors, none of which had clinically relevant effects on daratumumab PK.

We investigate knowledge retrieval with multi-modal queries, i.e. queries containing information split across image and text inputs, a challenging task that differs from previous work on cross-modal retrieval. We curate a new dataset called ReMuQ for benchmarking progress on this task. ReMuQ requires a system to retrieve knowledge from a large corpus by integrating contents from both text and image queries. We introduce a retriever model ``ReViz'' that can directly process input text and images to retrieve relevant knowledge in an end-to-end fashion without being dependent on intermediate modules such as object detectors or caption generators. We introduce a new pretraining task that is effective for learning knowledge retrieval with multimodal queries and also improves performance on downstream tasks. We demonstrate superior performance in retrieval on two datasets (ReMuQ and OK-VQA) under zero-shot settings as well as further improvements when finetuned on these datasets., Comment: ACL 2023

In-context learning (ICL), teaching a large language model (LLM) to perform a task with few-shot demonstrations rather than adjusting the model parameters, has emerged as a strong paradigm for using LLMs. While early studies primarily used a fixed or random set of demonstrations for all test queries, recent research suggests that retrieving semantically similar demonstrations to the input from a pool of available demonstrations results in better performance. This work expands the applicability of retrieval-based ICL approaches by demonstrating that even simple word-overlap similarity measures such as BM25 outperform randomly selected demonstrations. Furthermore, we extend the success of retrieval-based ICL to instruction-finetuned LLMs as well as Chain-of-Thought (CoT) prompting. For instruction-finetuned LLMs, we find that although a model has already seen the training data at training time, retrieving demonstrations from the training data at test time yields better results compared to using no demonstrations or random demonstrations. Last but not least, we train a task-specific demonstration retriever that outperforms off-the-shelf retrievers.

Pre-training on large corpora of text enables the language models to acquire a vast amount of factual and commonsense knowledge which allows them to achieve remarkable performance on a variety of language understanding tasks. They typically acquire this knowledge by learning from the pre-training text and capturing certain patterns from it. However, real-world settings often present scenarios that do not abide by these patterns i.e. scenarios that break the common assumptions. Can state-of-the-art NLP models correctly reason over the contexts of such scenarios? Addressing the above question, in this paper, we investigate the ability of models to correctly reason over contexts that break the common assumptions. To this end, we first systematically create evaluation data in which each data instance consists of (a) a common assumption, (b) a context that follows the assumption, (c) a context that breaks the assumption, and (d) questions based on the contexts. Then, through evaluations on multiple models including GPT-3 and Flan T5, we show that while doing fairly well on contexts that follow the common assumptions, the models struggle to correctly reason over contexts that break those assumptions. Specifically, the performance gap is as high as 20% absolute points. Furthermore, we thoroughly analyze these results revealing several interesting findings. We believe our work and findings will encourage and facilitate further research in developing more robust models that can also reliably reason over contexts that break the common assumptions. Data is available at \url{https://github.com/nrjvarshney/break_the_common_assumptions}., Comment: 6 pages

Logical reasoning is fundamental for humans yet presents a substantial challenge in the domain of Artificial Intelligence. Initially, researchers used Knowledge Representation and Reasoning (KR) systems that did not scale and required non-trivial manual effort. Recently, the emergence of large language models (LLMs) has demonstrated the ability to overcome various limitations of formal Knowledge Representation (KR) systems. Consequently, there's a growing interest in using LLMs for logical reasoning via natural language. This work strives to understand the proficiency of LLMs in logical reasoning by offering a brief review of the latest progress in this area; with a focus on the logical reasoning datasets, tasks, and the methods adopted to utilize LLMs for reasoning. To offer a thorough analysis, we have compiled a benchmark titled LogiGLUE. This includes 24 varied datasets encompassing deductive, abductive, and inductive reasoning. Utilizing LogiGLUE as a foundation, we have trained an instruction fine-tuned language model, resulting in LogiT5. We study single-task training, multi-task training, and "chain-of-thought" knowledge distillation fine-tuning technique to assess the performance of model across the different logical reasoning categories. We also assess various LLMs using LogiGLUE, and the findings indicate that LLMs excel most in abductive reasoning, followed by deductive reasoning, while they are least effective at inductive reasoning. We aim to shed light on the capabilities and potential pathways for enhancing logical reasoning proficiency in LLMs, paving the way for more advanced and nuanced developments in this critical field., Comment: Work in progress

Dialogue systems for Automatic Differential Diagnosis (ADD) have a wide range of real-life applications. These dialogue systems are promising for providing easy access and reducing medical costs. Building end-to-end ADD dialogue systems requires dialogue training datasets. However, to the best of our knowledge, there is no publicly available ADD dialogue dataset in English (although non-English datasets exist). Driven by this, we introduce MDDial, the first differential diagnosis dialogue dataset in English which can aid to build and evaluate end-to-end ADD dialogue systems. Additionally, earlier studies present the accuracy of diagnosis and symptoms either individually or as a combined weighted score. This method overlooks the connection between the symptoms and the diagnosis. We introduce a unified score for the ADD system that takes into account the interplay between symptoms and diagnosis. This score also indicates the system's reliability. To the end, we train two moderate-size of language models on MDDial. Our experiments suggest that while these language models can perform well on many natural language understanding tasks, including dialogue tasks in the general domain, they struggle to relate relevant symptoms and disease and thus have poor performance on MDDial. MDDial will be released publicly to aid the study of ADD dialogue research.

Atmospheric aerosols are ubiquitous in the atmosphere and can impact climate by scattering or absorbing solar radiation and acting as cloud condensation nuclei (CCN). Sea spray aerosols (SSA) are a major component of primary aerosols given the large ocean coverage of the Earth’s surface. SSA contains a variety of organic species, many of which are surface active, present throughout the ocean that can alter the climate relevant properties of SSA. Therefore, it is crucial to investigate the behavior and properties of these marine relevant species under different environmental conditions.In this dissertation study, the stability of different surface-active organics and biological species, such as palmitic acid, nonanoic acid, other lipids, and proteins at the air/water interface in the presence of a chemically complex underlying bulk phase is investigated. These different organic species are prominent in marine environments and are present within the sea surface microlayer (SSML), the topmost layer of the ocean, and on the surface of SSA. By using surface sensitive techniques such as surface pressure measurements and Infrared Reflectance Absorbance Spectroscopy (IRRAS), we have elucidated that palmitic acid monolayers on air/saltwater interfaces become less stable in the presence of light and a photosensitizer. Similarly, based on surface pressure measurements, model calculations and computational simulation, a thermodynamic cycle for nonanoic acid and its conjugated base between the air/water interface and the bulk phase is established. Furthermore, upon interaction with lipase, negatively charged, neutral and zwitterionic lipid monolayers were found to become disordered as demonstrated by surface pressure measurements, IRRAS measurements, water evaporation measurements, and computational simulations. Moreover, lipase is shown to interact differently depending on the charge of the lipid monolayer. The morphology of fatty acid monolayers with and without the presence o

The partitioning of medium-chain fatty acid surfactants such as nonanoic acid (NA) between the bulk phase and the air/water interface is of interest to a number of fields including marine and atmospheric chemistry. However, questions remain about the behavior of these molecules, the contributions of various relevant chemical equilibria, and the impact of pH, salt and bulk surfactant concentrations. In this study, the surface adsorption of nonanoic acid and its conjugate base is quantitatively investigated at various pH values, surfactant concentrations and the presence of salts. Surface concentrations of protonated and deprotonated species are dictated by surface-bulk equilibria which can be calculated from thermodynamic considerations. Notably we conclude that the surface dissociation constant of soluble surfactants cannot be directly obtained from these experimental measurements, however, we show that molecular dynamics (MD) simulation methods, such as free energy perturbation (FEP), can be used to calculate the surface acid dissociation constant relative to that in the bulk. These simulations show that nonanoic acid is less acidic at the surface compared to in the bulk solution with a pK a shift of 1.1 ± 0.6, yielding a predicted surface pK a of 5.9 ± 0.6. A thermodynamic cycle for nonanoic acid and its conjugate base between the air/water interface and the bulk phase can therefore be established. Furthermore, the effect of salts, namely NaCl, on the surface activity of protonated and deprotonated forms of nonanoic acid is also examined. Interestingly, salts cause both a decrease in the bulk pK a of nonanoic acid and a stabilization of both the protonated and deprotonated forms at the surface. Overall, these results suggest that the deprotonated medium-chain fatty acids under ocean conditions can also be present within the sea surface microlayer (SSML) present at the ocean/atmosphere interface due to the stabilization effect of the salts in the ocean. This allows

We have previously demonstrated that deletion of activin receptor-like kinase 1 (Alk1) or endoglin in a fraction of endothelial cells (ECs) induces brain arteriovenous malformations (bAVMs) in adult mice upon angiogenic stimulation. Here, we addressed three related questions: (1) could Alk1- mutant bone marrow (BM)-derived ECs (BMDECs) cause bAVMs? (2) is Alk1- ECs clonally expended during bAVM development? and (3) is the number of mutant ECs correlates to bAVM severity? For the first question, we transplanted BM from PdgfbiCreER;Alk12f/2f mice (EC-specific tamoxifen-inducible Cre with Alk1-floxed alleles) into wild-type mice, and then induced bAVMs by intra-brain injection of an adeno-associated viral vector expressing vascular endothelial growth factor and intra-peritoneal injection of tamoxifen. For the second question, clonal expansion was analyzed using PdgfbiCreER;Alk12f/2f;confetti+/- mice. For the third question, we titrated tamoxifen to limit Alk1 deletion and compared the severity of bAVM in mice treated with low and high tamoxifen doses. We found that wild-type mice with PdgfbiCreER;Alk12f/2f BM developed bAVMs upon VEGF stimulation and Alk1 gene deletion in BMDECs. We also observed clusters of ECs expressing the same confetti color within bAVMs and significant proliferation of Alk1- ECs at early stage of bAVM development, suggesting that Alk1- ECs clonally expanded by local proliferation. Tamoxifen dose titration revealed a direct correlation between the number of Alk1- ECs and the burden of dysplastic vessels in bAVMs. These results provide novel insights for the understanding of the mechanism by which a small fraction of Alk1 or endoglin mutant ECs contribute to development of bAVMs.

Recent state-of-the-art open-domain QA models are typically based on a two stage retriever-reader approach in which the retriever first finds the relevant knowledge/passages and the reader then leverages that to predict the answer. Prior work has shown that the performance of the reader usually tends to improve with the increase in the number of these passages. Thus, state-of-the-art models use a large number of passages (e.g. 100) for inference. While the reader in this approach achieves high prediction performance, its inference is computationally very expensive. We humans, on the other hand, use a more efficient strategy while answering: firstly, if we can confidently answer the question using our already acquired knowledge then we do not even use the external knowledge, and in the case when we do require external knowledge, we don't read the entire knowledge at once, instead, we only read that much knowledge that is sufficient to find the answer. Motivated by this procedure, we ask a research question "Can the open-domain QA reader utilize external knowledge efficiently like humans without sacrificing the prediction performance?" Driven by this question, we explore an approach that utilizes both 'closed-book' (leveraging knowledge already present in the model parameters) and 'open-book' inference (leveraging external knowledge). Furthermore, instead of using a large fixed number of passages for open-book inference, we dynamically read the external knowledge in multiple 'knowledge iterations'. Through comprehensive experiments on NQ and TriviaQA datasets, we demonstrate that this dynamic reading approach improves both the 'inference efficiency' and the 'prediction accuracy' of the reader. Comparing with the FiD reader, this approach matches its accuracy by utilizing just 18.32% of its reader inference cost and also outperforms it by achieving up to 55.10% accuracy on NQ Open., Comment: AAAI-23 Workshop on Knowledge Augmented Methods for NLP

The electrification of shared mobility has become popular across the globe. Many cities have their new shared e-mobility systems deployed, with continuously expanding coverage from central areas to the city edges. A key challenge in the operation of these systems is fleet rebalancing, i.e., how EVs should be repositioned to better satisfy future demand. This is particularly challenging in the context of expanding systems, because i) the range of the EVs is limited while charging time is typically long, which constrain the viable rebalancing operations; and ii) the EV stations in the system are dynamically changing, i.e., the legitimate targets for rebalancing operations can vary over time. We tackle these challenges by first investigating rich sets of data collected from a real-world shared e-mobility system for one year, analyzing the operation model, usage patterns and expansion dynamics of this new mobility mode. With the learned knowledge we design a high-fidelity simulator, which is able to abstract key operation details of EV sharing at fine granularity. Then we model the rebalancing task for shared e-mobility systems under continuous expansion as a Multi-Agent Reinforcement Learning (MARL) problem, which directly takes the range and charging properties of the EVs into account. We further propose a novel policy optimization approach with action cascading, which is able to cope with the expansion dynamics and solve the formulated MARL. We evaluate the proposed approach extensively, and experimental results show that our approach outperforms the state-of-the-art, offering significant performance gain in both satisfied demand and net revenue.

Hybrid retrievers can take advantage of both sparse and dense retrievers. Previous hybrid retrievers leverage indexing-heavy dense retrievers. In this work, we study "Is it possible to reduce the indexing memory of hybrid retrievers without sacrificing performance"? Driven by this question, we leverage an indexing-efficient dense retriever (i.e. DrBoost) and introduce a LITE retriever that further reduces the memory of DrBoost. LITE is jointly trained on contrastive learning and knowledge distillation from DrBoost. Then, we integrate BM25, a sparse retriever, with either LITE or DrBoost to form light hybrid retrievers. Our Hybrid-LITE retriever saves 13X memory while maintaining 98.0% performance of the hybrid retriever of BM25 and DPR. In addition, we study the generalization capacity of our light hybrid retrievers on out-of-domain dataset and a set of adversarial attacks datasets. Experiments showcase that light hybrid retrievers achieve better generalization performance than individual sparse and dense retrievers. Nevertheless, our analysis shows that there is a large room to improve the robustness of retrievers, suggesting a new research direction., Comment: accepted to ACL23

In this paper, we present the Multi-Forgery Detection Challenge held concurrently with the IEEE Computer Society Workshop on Biometrics at CVPR 2022. Our Multi-Forgery Detection Challenge aims to detect automatic image manipulations including but not limited to image editing, image synthesis, image generation, image photoshop, etc. Our challenge has attracted 674 teams from all over the world, with about 2000 valid result submission counts. We invited the Top 10 teams to present their solutions to the challenge, from which three teams are awarded prizes in the grand finale. In this paper, we present the solutions from the Top 3 teams, in order to boost the research work in the field of image forgery detection., Comment: Workshop and challenge summary paper, containing technical reports from different teams

Table Question Answering (TQA) is an important but under-explored task. Most of the existing QA datasets are in unstructured text format and only few of them use tables as the context. To the best of our knowledge, none of TQA datasets exist in the biomedical domain where tables are frequently used to present information. In this paper, we first curate a table question answering dataset, BioTABQA, using 22 templates and the context from a biomedical textbook on differential diagnosis. BioTABQA can not only be used to teach a model how to answer questions from tables but also evaluate how a model generalizes to unseen questions, an important scenario for biomedical applications. To achieve the generalization evaluation, we divide the templates into 17 training and 5 cross-task evaluations. Then, we develop two baselines using single and multi-tasks learning on BioTABQA. Furthermore, we explore instructional learning, a recent technique showing impressive generalizing performance. Experimental results show that our instruction-tuned model outperforms single and multi-task baselines on an average by ~23% and ~6% across various evaluation settings, and more importantly, instruction-tuned model outperforms baselines by ~5% on cross-tasks., Comment: BioASQ10 Workshop

Single-task models have proven pivotal in solving specific tasks; however, they have limitations in real-world applications where multi-tasking is necessary and domain shifts are exhibited. Recently, instructional prompts have shown significant improvement towards multi-task generalization; however, the effect of instructional prompts and Multi-Task Learning (MTL) has not been systematically studied in the biomedical domain. Motivated by this, this paper explores the impact of instructional prompts for biomedical MTL. We introduce the BoX, a collection of 32 instruction tasks for Biomedical NLP across (X) various categories. Using this meta-dataset, we propose a unified model termed In-BoXBART, that can jointly learn all tasks of the BoX without any task-specific modules. To the best of our knowledge, this is the first attempt to propose a unified model in the biomedical domain and use instructions to achieve generalization across several biomedical tasks. Experimental results indicate that the proposed model: 1) outperforms the single-task baseline by ~3% and multi-task (without instruction) baseline by ~18% on an average, and 2) shows ~23% improvement compared to the single-task baseline in few-shot learning (i.e., 32 instances per task) on an average. Our analysis indicates that there is significant room for improvement across tasks in the BoX, implying the scope for future research direction., Comment: NAACL 2022 Findings

The sequential recommendation aims at predicting the next items in user behaviors, which can be solved by characterizing item relationships in sequences. Due to the data sparsity and noise issues in sequences, a new self-supervised learning (SSL) paradigm is proposed to improve the performance, which employs contrastive learning between positive and negative views of sequences. However, existing methods all construct views by adopting augmentation from data perspectives, while we argue that 1) optimal data augmentation methods are hard to devise, 2) data augmentation methods destroy sequential correlations, and 3) data augmentation fails to incorporate comprehensive self-supervised signals. Therefore, we investigate the possibility of model augmentation to construct view pairs. We propose three levels of model augmentation methods: neuron masking, layer dropping, and encoder complementing. This work opens up a novel direction in constructing views for contrastive SSL. Experiments verify the efficacy of model augmentation for the SSL in the sequential recommendation. Code is available\footnote{\url{https://github.com/salesforce/SRMA}}., Comment: Preprint. Still under reivew

While both extractive and generative readers have been successfully applied to the Question Answering (QA) task, little attention has been paid toward the systematic comparison of them. Characterizing the strengths and weaknesses of the two readers is crucial not only for making a more informed reader selection in practice but also for developing a deeper understanding to foster further research on improving readers in a principled manner. Motivated by this goal, we make the first attempt to systematically study the comparison of extractive and generative readers for question answering. To be aligned with the state-of-the-art, we explore nine transformer-based large pre-trained language models (PrLMs) as backbone architectures. Furthermore, we organize our findings under two main categories: (1) keeping the architecture invariant, and (2) varying the underlying PrLMs. Among several interesting findings, it is important to highlight that (1) the generative readers perform better in long context QA, (2) the extractive readers perform better in short context while also showing better out-of-domain generalization, and (3) the encoder of encoder-decoder PrLMs (e.g., T5) turns out to be a strong extractive reader and outperforms the standard choice of encoder-only PrLMs (e.g., RoBERTa). We also study the effect of multi-task learning on the two types of readers varying the underlying PrLMs and perform qualitative and quantitative diagnosis to provide further insights into future directions in modeling better readers.

Data modification, either via additional training datasets, data augmentation, debiasing, and dataset filtering, has been proposed as an effective solution for generalizing to out-of-domain (OOD) inputs, in both natural language processing and computer vision literature. However, the effect of data modification on adversarial robustness remains unclear. In this work, we conduct a comprehensive study of common data modification strategies and evaluate not only their in-domain and OOD performance, but also their adversarial robustness (AR). We also present results on a two-dimensional synthetic dataset to visualize the effect of each method on the training distribution. This work serves as an empirical study towards understanding the relationship between generalizing to unseen domains and defending against adversarial perturbations. Our findings suggest that more data (either via additional datasets or data augmentation) benefits both OOD accuracy and AR. However, data filtering (previously shown to improve OOD accuracy on natural language inference) hurts OOD accuracy on other tasks such as question answering and image classification. We provide insights from our experiments to inform future work in this direction., Comment: ACL 2022 Findings

Information retrieval (IR) is essential in search engines and dialogue systems as well as natural language processing tasks such as open-domain question answering. IR serve an important function in the biomedical domain, where content and sources of scientific knowledge may evolve rapidly. Although neural retrievers have surpassed traditional IR approaches such as TF-IDF and BM25 in standard open-domain question answering tasks, they are still found lacking in the biomedical domain. In this paper, we seek to improve information retrieval (IR) using neural retrievers (NR) in the biomedical domain, and achieve this goal using a three-pronged approach. First, to tackle the relative lack of data in the biomedical domain, we propose a template-based question generation method that can be leveraged to train neural retriever models. Second, we develop two novel pre-training tasks that are closely aligned to the downstream task of information retrieval. Third, we introduce the ``Poly-DPR'' model which encodes each context into multiple context vectors. Extensive experiments and analysis on the BioASQ challenge suggest that our proposed method leads to large gains over existing neural approaches and beats BM25 in the small-corpus setting. We show that BM25 and our method can complement each other, and a simple hybrid model leads to further gains in the large corpus setting., Comment: Accepted at AAAI 2022

Over the past several years, the approximate message passing (AMP) algorithm has been applied to a broad range of problems, including compressed sensing (CS), robust regression, Bayesian estimation, etc. AMP was originally developed for compressed sensing based on the loopy belief propagation (BP). Compared to convex optimization based algorithms, AMP has low complexity and its performance can be rigorously characterized by a scalar state evolution (SE) in the case of a large independent and identically distributed (i.i.d.) (sub-) Gaussian matrix. AMP was then extended to solve general estimation problems with a generalized linear observation model. However, AMP performs poorly on a generic matrix such as non-zero mean, rank-deficient, correlated, or ill-conditioned matrix, resulting in divergence and degraded performance. It was discovered later that applying AMP to a unitary transform of the original model can remarkably enhance the robustness to difficult matrices. This variant is named unitary AMP (UAMP), or formally called UTAMP. In this thesis, leveraging UAMP, we propose UAMP-SBL for sparse signal recovery and Bi-UAMP for bilinear recovery, both of which inherit the low complexity and robustness of UAMP. Sparse Bayesian learning (SBL) is a powerful tool for recovering a sparse signal from noisy measurements, which finds numerous applications in various areas. As a traditional implementation of SBL, e.g., Tipping’s method, involves matrix inversion in each iteration, the computational complexity can be prohibitive for large scale problems. To circumvent this, AMP and its variants have been used as low-complexity solutions. Unfortunately, they will diverge for ‘difficult’ measurement matrices as previously mentioned. In this thesis, leveraging UAMP, a novel SBL algorithm called UAMP-SBL is proposed where UAMP is incorporated into the structured variational message passing (SVMP) to handle the most computationally intensive part of message computations. It is sh

Information Retriever (IR) aims to find the relevant documents (e.g. snippets, passages, and articles) to a given query at large scale. IR plays an important role in many tasks such as open domain question answering and dialogue systems, where external knowledge is needed. In the past, searching algorithms based on term matching have been widely used. Recently, neural-based algorithms (termed as neural retrievers) have gained more attention which can mitigate the limitations of traditional methods. Regardless of the success achieved by neural retrievers, they still face many challenges, e.g. suffering from a small amount of training data and failing to answer simple entity-centric questions. Furthermore, most of the existing neural retrievers are developed for pure-text query. This prevents them from handling multi-modality queries (i.e. the query is composed of textual description and images). This proposal has two goals. First, we introduce methods to address the abovementioned issues of neural retrievers from three angles, new model architectures, IR-oriented pretraining tasks, and generating large scale training data. Second, we identify the future research direction and propose potential corresponding solution., Comment: Accepted to NAACL 2022 SRW

Shared e-mobility services have been widely tested and piloted in cities across the globe, and already woven into the fabric of modern urban planning. This paper studies a practical yet important problem in those systems: how to deploy and manage their infrastructure across space and time, so that the services are ubiquitous to the users while sustainable in profitability. However, in real-world systems evaluating the performance of different deployment strategies and then finding the optimal plan is prohibitively expensive, as it is often infeasible to conduct many iterations of trial-and-error. We tackle this by designing a high-fidelity simulation environment, which abstracts the key operation details of the shared e-mobility systems at fine-granularity, and is calibrated using data collected from the real-world. This allows us to try out arbitrary deployment plans to learn the optimal given specific context, before actually implementing any in the real-world systems. In particular, we propose a novel multi-agent neural search approach, in which we design a hierarchical controller to produce tentative deployment plans. The generated deployment plans are then tested using a multi-simulation paradigm, i.e., evaluated in parallel, where the results are used to train the controller with deep reinforcement learning. With this closed loop, the controller can be steered to have higher probability of generating better deployment plans in future iterations. The proposed approach has been evaluated extensively in our simulation environment, and experimental results show that it outperforms baselines e.g., human knowledge, and state-of-the-art heuristic-based optimization approaches in both service coverage and net revenue.

Zhili Zhang,1,* Hongli Yang,2,* Man Luo3 1Department of Surgical, Wuhan Third Hospital Affiliated to Wuhan University, Wuhan, 430070, People’s Republic of China; 2Department of Public Health, The First Community Health Service Center of Guanshan, Wuhan, 430073, People’s Republic of China; 3Department of Nursing, Wuhan Third Hospital Affiliated to Wuhan University, Wuhan, 430070, People’s Republic of China*These authors contributed equally to this workCorrespondence: Man LuoDepartment of Nursing, Wuhan Third Hospital Affiliated to Wuhan University, Wuhan, 430070, People’s Republic of ChinaTel +86 18207139227Fax +86 27-88850381Email lm229603580@163.comPurpose: To explore the correlation between community-acquired pressure injury (CAPI) and comorbidities in elderly patients with emergency admission.Patients and Methods: Patients aged 65 years or above were enrolled from multiple departments, such as Internal Medicine, Surgery, Geriatrics, and Intensive Care Unit of Wuhan Third Hospital, which is affiliated to Wuhan University, from January to December 2020. Comorbidity data were extracted using the 10th edition of the International Classification of Diseases (ICD-10) from the hospital electronic medical record system, and the Charlson Comorbidity Index (CCI) was calculated using these data. Participants were divided into two groups according to whether pressure injury was present at admission. The baseline characteristics of the two groups were compared using Student’s t-tests, Mann–Whitney U-tests, and chi-square tests. Univariate and multivariate logistic regression models were constructed to explore the relationship between CAPI and the CCI. Smooth curve fitting was used to show the relationship between the CCI and CAPI. By drawing the receiver operating characteristic curve, the CCI was used to predict CAPI.Results: A total of 5759 participants with an average age of 75.1 ± 7.6 were included in this population-based study. The p

We report the population pharmacokinetic (PK) and exposure-response analyses of a novel subcutaneous formulation of daratumumab (DARA) using data from 3 DARA subcutaneous monotherapy studies (PAVO Part 2, MMY1008, COLUMBA) and 1 combination therapy study (PLEIADES). Results were based on 5159 PK samples from 742 patients (DARA 1800 mg subcutaneously, n = 487 [monotherapy, n = 288; combination therapy, n = 199]; DARA 16 mg/kg intravenously, n = 255 [all monotherapy, in COLUMBA]; age, 33-92 years; weight, 28.6-147.6 kg). Subcutaneous and intravenous DARA monotherapies were administered once every week for cycles 1-2, once every 2 weeks for cycles 3-6, and once every 4 weeks thereafter (1 cycle is 28 days). The subcutaneous DARA combination therapy was administered with the adaptation of corresponding standard-of-care regimens. PK samples were collected between cycle 1 and cycle 12. Among monotherapy studies, throughout the treatment period, subcutaneous DARA provided similar/slightly higher trough concentrations (Ctrough) versus intravenous DARA, with lower maximum concentrations and smaller peak-to-trough fluctuations. The PK profile was consistent between subcutaneous DARA monotherapy and combination therapies. The exposure-response relationship between daratumumab PK and efficacy or safety end points was similar for subcutaneous and intravenous DARA. Although the ≤65-kg subgroup reported a higher incidence of neutropenia, no relationship was found between the incidence of neutropenia and exposure, which was attributed, in part, to the preexisting imbalance in neutropenia between subcutaneous DARA (45.5%) and intravenous DARA (19%) in patients ≤50 kg. A flat relationship was observed between body weight and any grade and at least grade 3 infections. The results support the DARA 1800-mg subcutaneous flat dose as an alternative to the approved intravenous DARA 16 mg/kg.

Li Y, Song B, Li Z, et al. Nat Sci Sleep. 2021;13:821–828. On page 823, Results, The Comparison of Demographic Characteristics of the Two Groups section, line 3, the text “duration of the operation (min) (P = 0.062)” should read “duration of surgery (min) (P = 0.517)”. Table 1 on page 824, the Duration of surgery (min) row is incorrect. The correct Table 1 is shown in Download Article. Read the original article

In the open book question answering (OBQA) task, selecting the relevant passages and sentences from distracting information is crucial to reason the answer to a question. HotpotQA dataset is designed to teach and evaluate systems to do both passage ranking and sentence selection. Many existing frameworks use separate models to select relevant passages and sentences respectively. Such systems not only have high complexity in terms of the parameters of models but also fail to take the advantage of training these two tasks together since one task can be beneficial for the other one. In this work, we present a simple yet effective framework to address these limitations by jointly ranking passages and selecting sentences. Furthermore, we propose consistency and similarity constraints to promote the correlation and interaction between passage ranking and sentence selection.The experiments demonstrate that our framework can achieve competitive results with previous systems and outperform the baseline by 28\% in terms of exact matching of relevant sentences on the HotpotQA dataset., Comment: Accepted to NAACL SWR 2022

Knowledge-based visual question answering (VQA) requires answering questions with external knowledge in addition to the content of images. One dataset that is mostly used in evaluating knowledge-based VQA is OK-VQA, but it lacks a gold standard knowledge corpus for retrieval. Existing work leverage different knowledge bases (e.g., ConceptNet and Wikipedia) to obtain external knowledge. Because of varying knowledge bases, it is hard to fairly compare models' performance. To address this issue, we collect a natural language knowledge base that can be used for any VQA system. Moreover, we propose a Visual Retriever-Reader pipeline to approach knowledge-based VQA. The visual retriever aims to retrieve relevant knowledge, and the visual reader seeks to predict answers based on given knowledge. We introduce various ways to retrieve knowledge using text and images and two reader styles: classification and extraction. Both the retriever and reader are trained with weak supervision. Our experimental results show that a good retriever can significantly improve the reader's performance on the OK-VQA challenge. The code and corpus are provided in https://github.com/luomancs/retriever\_reader\_for\_okvqa.git, Comment: accepted at EMNLP 2021

We propose a deep signature/log-signature FBSDE algorithm to solve forward-backward stochastic differential equations (FBSDEs) with state and path dependent features. By incorporating the deep signature/log-signature transformation into the recurrent neural network (RNN) model, our algorithm shortens the training time, improves the accuracy, and extends the time horizon comparing to methods in the existing literature. Moreover, our algorithms can be applied to a wide range of applications such as state and path dependent option pricing involving high-frequency data, model ambiguity, and stochastic games, which are linked to parabolic partial differential equations (PDEs), and path-dependent PDEs (PPDEs). Lastly, we also derive the convergence analysis of the deep signature/log-signature FBSDE algorithm.

GQA~\citep{hudson2019gqa} is a dataset for real-world visual reasoning and compositional question answering. We found that many answers predicted by the best vision-language models on the GQA dataset do not match the ground-truth answer but still are semantically meaningful and correct in the given context. In fact, this is the case with most existing visual question answering (VQA) datasets where they assume only one ground-truth answer for each question. We propose Alternative Answer Sets (AAS) of ground-truth answers to address this limitation, which is created automatically using off-the-shelf NLP tools. We introduce a semantic metric based on AAS and modify top VQA solvers to support multiple plausible answers for a question. We implement this approach on the GQA dataset and show the performance improvements. Code and data are available in this link \url{https://github.com/luomancs/alternative_answer_set.git}., Comment: accepted to EACL 2021

As facial interaction systems are prevalently deployed, security and reliability of these systems become a critical issue, with substantial research efforts devoted. Among them, face anti-spoofing emerges as an important area, whose objective is to identify whether a presented face is live or spoof. Recently, a large-scale face anti-spoofing dataset, CelebA-Spoof which comprised of 625,537 pictures of 10,177 subjects has been released. It is the largest face anti-spoofing dataset in terms of the numbers of the data and the subjects. This paper reports methods and results in the CelebA-Spoof Challenge 2020 on Face AntiSpoofing which employs the CelebA-Spoof dataset. The model evaluation is conducted online on the hidden test set. A total of 134 participants registered for the competition, and 19 teams made valid submissions. We will analyze the top ranked solutions and present some discussion on future work directions., Comment: Technical report. Challenge website: https://competitions.codalab.org/competitions/26210

Yang Li, 1,* Bijia Song, 2,* Zhen Li, 3 Jingjie Wan, 4 Man Luo, 1 Wenxin Wei, 1 Junchao Zhu 1 1Department of Anesthesiology, Shengjing Hospital of China Medical University, Shenyang, Liaoning, People’s Republic of China; 2Department of Anesthesiology, Friendship Hospital of Capital Medical University, Beijing, People’s Republic of China; 3Department of Anesthesiology and Perioperative Medicine, Shanghai Fourth People’s Hospital Affiliated to Tongji University School of Medicine, Shanghai, People’s Republic of China; 4Department of Anesthesiology, First Affiliated Hospital of Suzhou University, Suzhou, People’s Republic of China*These authors contributed equally to this workCorrespondence: Junchao ZhuDepartment of Anesthesiology, Shengjing Hospital of China Medical University, Shenyang, Liaoning, People’s Republic of ChinaTel +86 18940257257Email zhujunchao1@hotmail.comObjective: The primary purpose was to compare the effects of sufentanil and fentanyl on the postoperative sleep quality. And the secondary purposes were to evaluate perioperative hemodynamics, postoperative pain and complications of children undergoing tonsillectomy and adenotomy.Methods: Seventy-six patients were randomly assigned to the sufentanil or fentanyl group. The subjective sleep quality was assessed by the Athens Insomnia Scale (AIS) on the night before surgery (Sleep preop 1), the first night after surgery (Sleep POD 1), and the third night after surgery (Sleep POD 3). The Faces Pain Scale-Revised (FPS-R) was used to evaluate the postoperative pain level 24 hours after surgery. The Observer’s Assessment of Alertness and Sedation (OAA/S) scale was used to assess the level of sedation in children. Perioperative hemodynamics and adverse effects were also evaluated.Results: The AIS score in the sufentanil group was significantly lower at Sleep POD 1 and Sleep POD 3 (P < 0.001, respectively). Children in the sufentanil g

Recently, several promising approximate message passing (AMP) based algorithms have been developed for bilinear recovery with model $\boldsymbol{Y}=\sum_{k=1}^K b_k \boldsymbol{A}_k \boldsymbol{C} +\boldsymbol{W} $, where $\{b_k\}$ and $\boldsymbol{C}$ are jointly recovered with known $\boldsymbol{A}_k$ from the noisy measurements $\boldsymbol{Y}$. The bilinear recover problem has many applications such as dictionary learning, self-calibration, compressive sensing with matrix uncertainty, etc. In this work, we propose a new bilinear recovery algorithm based on AMP with unitary transformation. It is shown that, compared to the state-of-the-art message passing based algorithms, the proposed algorithm is much more robust and faster, leading to remarkably better performance., Comment: 11 pages, 12 figures

A recent work has shown that transformers are able to "reason" with facts and rules in a limited setting where the rules are natural language expressions of conjunctions of conditions implying a conclusion. Since this suggests that transformers may be used for reasoning with knowledge given in natural language, we do a rigorous evaluation of this with respect to a common form of knowledge and its corresponding reasoning -- the reasoning about effects of actions. Reasoning about action and change has been a top focus in the knowledge representation subfield of AI from the early days of AI and more recently it has been a highlight aspect in common sense question answering. We consider four action domains (Blocks World, Logistics, Dock-Worker-Robots and a Generic Domain) in natural language and create QA datasets that involve reasoning about the effects of actions in these domains. We investigate the ability of transformers to (a) learn to reason in these domains and (b) transfer that learning from the generic domains to the other domains.

Man Luo,1,* Bijia Song,1,2,* Junchao Zhu1 1Department of Anesthesiology, Shengjing Hospital of China Medical University, Shenyang, Liaoning, People’s Republic of China; 2Department of Anesthesiology, Friendship Hospital of Capital Medical University, Beijing, People’s Republic of China*These authors contributed equally to this workCorrespondence: Junchao Zhu Department of AnesthesiologyShengjing Hospital of China Medical University, Shenyang, Liaoning, People’s Republic of ChinaEmail zhujunchao1@hotmail.comAbstract: General anesthesia produces a state of drug-induced unconsciousness that is controlled by the extent and duration of administered agents. Whether inhalation or intravenous in formulation, such agents may interfere with normal sleep–wake cycles, impairing postoperative sleep quality and creating complications. Electroacupuncture is a new approach widely applied in clinical practice during recent years. This particular technology helps regulate neurotransmitter concentrations in the brain, lowering norepinephrine and dopamine levels to improve sleep quality. It also alleviates surgical pain that degrades postoperative sleep quality after general anesthesia by downregulating immune activity (SP, NK-1, and COX-1) and upregulating serotonin receptor (5-HT1AR, 5-HT2AR) and endocannabinoid expression levels. However, large-scale, multicenter studies are still needed to determine the optimal duration, frequency, and timing of electroacupuncture for such use.Keywords: general anesthesia, sleep quality, acupoints, electroacupuncture

Vibrational sum frequency generation (VSFG) spectroscopy and surface pressure measurements are used to investigate the adsorption of a globular protein, bovine serum albumin (BSA), at the air/water interface with and without the presence of salts. We find at low (2 to 5 ppm) protein concentrations, which is relevant to environmental conditions, both VSFG and surface pressure measurements of BSA behave drastically different from at higher concentrations. Instead of emerging to the surface immediately, as observed at 1000 ppm, protein adsorption kinetics is on the order of tens of minutes at lower concentrations. Most importantly, salts strongly enhance the presence of BSA at the interface. This "salting up" effect differs from the well-known "salting out" effect as it occurs at protein concentrations well-below where "salting out" occurs. The dependence on salt concentration suggests this effect relates to a large extent electrostatic interactions and volume exclusion. Additionally, results from other proteins and the pH dependence of the kinetics indicate that salting up depends on the flexibility of proteins. This initial report demonstrates "salting up" as a new type of salt-driven interfacial phenomenon, which is worthy of continued investigation given the importance of salts in biological and environmental aqueous systems.

Ting Guo,1–3 Wei Guo,1–3 Min Song,1–3 Shanshan Ni,1–3 Man Luo,1–3 Ping Chen,1–3 Hong Peng1–3 1Department of Respiratory and Critical Care Medicine, The Second Xiangya Hospital of Central-South University, Changsha, Hunan 410011, People’s Republic of China; 2Research Unit of Respiratory Disease, Central-South University, Changsha, Hunan 410011, People’s Republic of China; 3The Respiratory Disease Diagnosis and Treatment Center of Hunan Province, Changsha, Hunan 410011, People’s Republic of ChinaCorrespondence: Hong PengDepartment of Respiratory and Critical Care Medicine, The Second Xiangya Hospital Of Central-South University, No. 139 Renmin Middle Road, Changsha, Hunan 410011, People’s Republic of ChinaTel +86-13807494486Fax +86 731-85295147Email penghong66@csu.edu.cnIntroduction: Paradoxical reaction refers to deterioration of the original tuberculosis lesions or emergence of new infiltrative lesions during anti-tuberculosis treatment. The common manifestations of paradoxical reaction include new pleural effusion, cerebral tuberculosis and lymphadenitis. Paradoxical reaction manifested by new pulmonary mass is rare.Patients and methods: This article summarizes and analyzes the clinical manifestations, chest CT, laboratory findings, treatments, pathological biopsy results of five patients diagnosed as paradoxical reaction in the form of new pulmonary mass. A literature review related to paradoxical reaction was conducted.Results: Five patients diagnosed as pulmonary tuberculosis or tuberculous pleuritis received systematic anti-tuberculosis treatmensssts. New pulmonary masses were found by CT scans during the follow-ups. The patients were negative for tumor markers, examination of rheumatoid connective tissue disease and G/GM test. The original anti-tuberculosis treatments were continued. All of the masses were diminished gradually.Conclusion: Paradoxical r

LPMLN is a probabilistic extension of answer set programs with the weight scheme adapted from Markov Logic. We study the concept of strong equivalence in LPMLN, which is a useful mathematical tool for simplifying a part of an LPMLN program without looking at the rest of it. We show that the verification of strong equivalence in LPMLN can be reduced to equivalence checking in classical logic via a reduct and choice rules as well as to equivalence checking under the "soft" logic of here-and-there. The result allows us to leverage an answer set solver for LPMLN strong equivalence checking. The study also suggests us a few reformulations of the LPMLN semantics using choice rules, the logic of here-and-there, and classical logic., Comment: In Proceedings ICLP 2019, arXiv:1909.07646. arXiv admin note: text overlap with arXiv:1905.07550

Sparse Bayesian learning (SBL) can be implemented with low complexity based on the approximate message passing (AMP) algorithm. However, it is vulnerable to `difficult' measurement matrices as AMP can easily diverge. Damped AMP has been used to alleviate the problem at the cost of slowing the convergence speed. In this work, we propose an SBL algorithm based on the AMP with unitary transformation (UTAMP), where the shape parameter of the hyperprior is tuned automatically. It is shown that, compared to the state-of-the-art AMP based SBL algorithm, the proposed UTAMP-SBL is much more robust and much faster, leading to remarkably better performance. It is shown that in many cases, UTAMP-SBL can approach the support-oracle bound closely., Comment: Part of our work in progress (presented in IEEE APWCS 2019)

Strong equivalence is a well-studied and important concept in answer set programming (ASP). $\rm LP^{MLN}$ is a probabilistic extension of answer set programs with the weight scheme adapted from Markov Logic. Because of the semantic differences, strong equivalence for ASP does not simply carry over to $\rm LP^{MLN}$. I study the concept of strong equivalence in $\rm LP^{MLN}$ with the goal of extending strong equivalence to $\rm LP^{MLN}$ programs. My study shows that the verification of strong equivalence in $\rm LP^{MLN}$ can be reduced to equivalence checking in classical logic plus weight consideration.The result allows us to leverage an answer set solver for checking strong equivalence in $\rm LP^{MLN}$. Furthermore, this study also suggests us a few reformulations of the $\rm LP^{MLN}$ semantics using choice rules, logic of here and there, and the second-order logic. I will present my work result of strong equivalence for $\rm LP^{MLN}$ and talk about my next steps for research: one is approximately strong equivalence, and another is the integration of fuzzy logic with neural network.