Pay more attention to the robustness of LLMs on adversarial prompt for instruction data mining.

Instruction tuning has emerged as a paramount method for tailoring the behaviors of LLMs. Recent studies have unveiled the potential for LLMs to achieve high performance through fine-tuning with a limited quantity of high-quality instruction data. Instruction-Following Difficulty is one of the most representative approaches in instruction data mining, which involves selecting samples where LLMs fail to generate response that align with the provided instructions as the high-quality instruction data.

Co-design protein sequence and structure in discrete space via generative flow.

Motivation: Generative models have demonstrated considerable promise in de novo protein design. Traditional approaches typically focus on either sequence or structure in isolation, limiting the capacity to explore the intricate sequence-structure landscape and achieve optimal designs. However, joint protein sequence and structure co-design remains a largely underexplored challenge.

Phosphate-coupled high-carbon ferromanganese particles synergistically regulate co-composting of seaweed and corn starch residue: Improving nitrogen cycling and accelerating humification.

A substantial amount of seaweed waste, rich in plant-stimulating properties, holds great potential for composting. However, without proper catalysts or additives, composting and humification are slow and inefficient, causing carbon and nitrogen losses. This study explores a novel approach to enhance composting efficacy by coupling phosphate with high-carbon ferromanganese particles (HCFMP).

Exploring Structure Diversity in Atomic Resolution Microscopy With Graph.

The emergence of deep learning (DL) has provided great opportunities for the high-throughput analysis of atomic-resolution micrographs. However, the DL models trained by image patches in fixed size generally lack efficiency and flexibility when processing micrographs containing diversified atomic configurations. Herein, inspired by the similarity between the atomic structures and graphs, a few-shot learning framework based on an equivariant graph neural network (EGNN) to analyze a library of atomic structures (e.

Colorimetric, Quantitative, and Portable Liquid Crystal Elastomer Biosensing of Cholesterol and Malathion.

Cholesteric liquid crystal elastomers (CLCEs) possess a unique characteristic that allows the manipulation of the reflected color by adjusting the spacing between layers. This attribute can create a reliable and cost-effective colorimetric sensing platform for the on-site detection of various substances. In this study, CLCE films were coupled with cross-linked poly(acrylic acid) (CLCE-PAA) or poly(dimethylaminoethyl methacrylate) (CLCE-PDA) films to monitor cholesterol and malathion levels, respectively.

Fluorine-Free Ion-Selective Membrane with Enhanced Mg Transport for Mg-Organic Batteries.

Magnesium batteries offer a safer alternative for next-generation battery technology due to their insusceptibility to dendrite deposition. Selective membranes tailored for magnesium-ion conduction will unlock further technological advancement. Herein, we demonstrate fluorine-free magnesiated sulfonated poly(ether ether ketone) (Mg-SPEEK) selective membranes capable of facilitating magnesium-ion conduction while effectively rejecting soluble organic species.

Single Na- and K-Ion-Conducting Sulfonated -NH-Linked Covalent Organic Frameworks.

Highly ion-conductive solid electrolytes of nonlithium ions (sodium or potassium ions) are necessary for pursuing a more cost-effective and sustainable energy storage. Here, two classes of sulfonated -NH-linked covalent organic frameworks (COFs), specifically designed for sodium or potassium ion conduction (named i-COF-2 (Na or K) and i-COF-3 (Na or K)), were synthesized through a straightforward, one-step process using affordable starting materials. Remarkably, these COFs demonstrate high ionic conductivity at room temperature─3.

Vertical distribution and influencing factors of soil PAHs under different ecosystem habitats in the Liaohe River Estuary Wetlands, Northeastern China.

The vertical distribution, sources and influencing factors of polycyclic aromatic hydrocarbons (PAHs) in soil across ecosystem habitats were investigated around the Liaohe River Estuary (LRE) Wetland. The concentration of ƩPAHs ranged from 41.0 to 435.

Two-Dimensional Electrically Conductive Metal-Organic Framework Boosts Synaptic Plasticity for Dynamic Image Refresh, Classification, and Efferent Neuromuscular Systems.

We present a two-dimensional (2D) electrically conductive metal-organic framework (EC-MOF)-based artificial synapse. The intrinsic electronic conductivity and subnanometer channels of the EC-MOF facilitate efficient ion diffusion, enable a high density of active redox centers, and significantly enhance capacitance within the artificial synapse. As a result, the synapse operates at an ultralow voltage of 10 mV and exhibits a remarkably low power consumption of approximately 1 fW, along with the longest retention time recorded for two-terminal electrolyte-type artificial synapses to date.

Ion transport mechanisms in covalent organic frameworks: implications for technology.

Covalent organic frameworks (COFs) have emerged as promising materials for ion conduction due to their highly tunable structures and excellent electrochemical stability. This review paper explores the mechanisms of ion conduction in COFs, focusing on how these materials facilitate ion transport across their ordered structures, which is crucial for applications such as solid electrolytes in batteries and fuel cells. We discuss the design strategies employed to enhance ion conductivity, including pore size optimization, functionalization with ionic groups, and the incorporation of solvent molecules and salts.

Less confidence, less forgetting: Learning with a humbler teacher in exemplar-free Class-Incremental learning.

Class-Incremental learning (CIL) is challenging due to catastrophic forgetting (CF), which escalates in exemplar-free scenarios. To mitigate CF, Knowledge Distillation (KD), which leverages old models as teacher models, has been widely employed in CIL. However, based on a case study, our investigation reveals that the teacher model exhibits over-confidence in unseen new samples.

Masked Modeling-Based Ultrasound Image Classification via Self-Supervised Learning.

Recently, deep learning-based methods have emerged as the preferred approach for ultrasound data analysis. However, these methods often require large-scale annotated datasets for training deep models, which are not readily available in practical scenarios. Additionally, the presence of speckle noise and other imaging artifacts can introduce numerous hard examples for ultrasound data classification.

Changes in the labor share of enterprises under the green and low-carbon transition: Empirical evidence from China's carbon emission trading pilot.

Green and low carbon reflect the high-quality development, while income distribution is an indicator of the balance of development. Is there a lack of fairness in the process of green and low carbon transition of enterprises? Using data from A-share listed companies from 2009 to 2016, this paper constructs a DID identification framework for controlling the endogeneity problem using the 2013 carbon trading policy pilot as a quasi-natural experiment to empirically test the impact of corporate low-carbon transformation on corporate labor income share in the context of carbon trading policy. The findings indicate that carbon trading policy decreases the labor income share of firms.

Exploring the Knowledge of an Outstanding Protein to Protein Interaction Transformer.

Protein-to-protein interaction (PPI) prediction aims to predict whether two given proteins interact or not. Compared with traditional experimental methods of high cost and low efficiency, the current deep learning based approach makes it possible to discover massive potential PPIs from large-scale databases. However, deep PPI prediction models perform poorly on unseen species, as their proteins are not in the training set.

Poly(2-vinylpyridine)/MCM-41 Composites with Micropores and Switchable Mesopores for the Removal of Cr(VI).

Porous structure design and reversible regulation of pore size during adsorption-desorption are crucial to the removal of pollutants in water such as Cr(VI). In this paper, micropores and switchable mesopores were constructed on MCM-41 to further improve adsorption-desorption performance of Cr(VI) via the confinement effect of micropores and opening and closing of mesopores. 2-Vinylpyridine was introduced and polymerized into the pores and on the pore mouth of MCM41 modified by C═C group (AM41) under the irradiation of ultraviolet light.

Surgical repair of post myocardial infarction ventricular septal defect: a retrospective analysis of a single institution experience.

Introduction: Ventricular septal defect (VSD) is a mechanical complication of acute myocardial infarction (MI) with a very high mortality, despite advances in surgical and circulatory support. The tremendous hemodynamic disturbance and the severely fragile myocardium render surgical repair a great challenge. The optimal time of surgical repair with or without circulatory support is still controversial.

Effects of oyster shells on maturity and calcium activation in organic solid waste compost.

With the rapid development of aquaculture, the production of oyster shells has surged, posing a potential threat to the environment. While oyster shell powder is widely recognized for its inherent alkalinity and rich calcium carbonate content, making it a superior soil conditioner, its role in organic solid waste composting remains underexplored. To investigate the effects of varying concentrations of oyster shell powder on compost maturation and calcium activation, this study employed thermophilic co-composting with acidic sugar residue and bean pulp, incorporating 0% (control), 10% (T1), 20% (T2), 30% (T3), and 40% (T4) oyster shell powder.

The impact of household wealth gap on individual's mental health.

Background: Improving the individual's mental health is important for sustainable economic and social development. Although some studies found that household wealth gap may affect individuals' mental health, few studies have clarified the causal relationship between household wealth gap between mental health in China. This study examines the impact of the household wealth gap on individuals' mental health using data from the 2012-2018 China Family Panel Survey.

Self-supervised learning-based underwater acoustical signal classification via mask modeling.

The classification of underwater acoustic signals has garnered a great deal of attention in recent years due to its potential applications in military and civilian contexts. While deep neural networks have emerged as the preferred method for this task, the representation of the signals plays a crucial role in determining the performance of the classification. However, the representation of underwater acoustic signals remains an under-explored area.