Gefitinib is a first-line therapy in the EGFR-mutated nonsmall cell lung cancer (NSCLC). However, development of drug resistance almost unavoidable, thus leading to an unsustainable regimen. EGFRT790M mutation major cause responsible for molecular-targeting failure NSCLC. Although recently approved osimertinib effective EGFRT790M-positive NSCLC, osimertinib-resistant EGFR rapidly developed, too. In this study, we proposed tumor-associated macrophage (TAM) reprogramming strategy overcoming...

Background Glioblastoma (GBM) treatment is undermined by the suppressive tumor immune microenvironment (TIME). Seek for effective methods brain TIME modulation a pressing need. However, there are two major challenges against achieving goal: first, to screen drugs with TIME-remodeling functions and, second, develop targeting system delivering drugs. Methods In this study, an α7 nicotinic acetylcholine receptors (nAChRs)-binding peptide D CDX was used modify codelivery liposomes achieve...

Effective treatment of metastatic (stage IV) breast cancers remains a formidable challenge. To address this issue, cell-penetrating peptide-assisted liposomal system was developed for codelivery doxorubicin and andrographolide. This nanomedicine-based combination therapy showed the ability to inhibit in vitro migration invasion 4T1 cells through wound healing transwell assays. Furthermore, delivery exhibited enhanced accumulation tumor tissues deep intratumoral penetration. The synergistic...

Aspect sentiment quad prediction (ASQP) aims to predict the elements for a given sentence, which is critical task in field of aspect-based analysis. However, data imbalance issue has not received sufficient attention ASQP task. In this paper, we divide into two-folds, quad-pattern and aspect-category imbalance, propose an Adaptive Data Augmentation (ADA) framework tackle issue. Specifically, augmentation process with condition function adaptively enhances tail patterns aspect categories,...

In recent years, temporal knowledge graph (TKG) reasoning has received significant attention. Most existing methods assume that all timestamps and corresponding graphs are available during training, which makes it difficult to predict future events. To address this issue, works learn infer events based on historical information. However, these do not comprehensively consider the latent patterns behind changes, pass information selectively, update representations appropriately accurately....

Effective delivery is the primary barrier against clinical translation of gene therapy. Yet there remains too much unknown in mechanisms, even for most investigated polymeric carrier (i.e., PEI). As a consequence, conflicting results have been often seen literature due to large variability experimental conditions and operations. Therefore, some key parameters should be identified thus strictly controlled formulation process. Methods: The effect processing (e.g., concentration or mixture...

A mannosylated, bioreducible Man-PEI_5k–CPP/pTRAIL system was developed for treating colon cancer.

Objective quality assessment of stereoscopic panoramic images becomes a challenging problem owing to the rapid growth 360-degree contents. Different from traditional 2D image (IQA), more complex aspects are involved in 3D omnidirectional IQA, especially unlimited field view (FoV) and extra depth perception, which brings difficulty evaluate experience (QoE) images. In this paper, we propose multi-viewport based full-reference stereo 360 IQA model. Due freely changeable viewports when browsing...

Drug efflux-associated multidrug resistance (MDR) is a main obstacle to effective cancer chemotherapy. Large molecule drugs are not the substrates of P-glycoprotein, and can circumvent drug efflux be retained inside cells. In this article we report polymer-drug conjugate nanoparticulate system that overcome MDR based on size-related exclusion effect.Doxorubicin was coupled with triblock polymeric material cell-penetrating TAT-PEG-poly(aspartic acid). The amphiphilic macromolecules (termed...

Cross-domain recommendation (CDR) has been proven as a promising way to tackle the user cold-start problem, which aims make recommendations for users in target domain by transferring preference derived from source domain. Traditional CDR studies follow embedding and mapping (EMCDR) paradigm, transfers representations learning user-shared function, neglecting user-specific preference. Recent attempt learn functions meta-learning regards each user's an individual task, but neglects...

It is crucial to explore a suitable landslide susceptibility model with an excellent prediction capability for rapid evaluation and disaster relief in seismic regions different lithological features. In this study, we selected two typical events, the Jiuzhaigou Minxian earthquakes, which occurred Alpine karst loess regions, respectively. Eight influencing factors five models were chosen calculate of landslide, including information (I) model, certainty factor (CF) logistic regression (LR) I...

Natural language (NL) based vehicle retrieval aims to search specific given text description. Different from the image-based retrieval, NL-based requires considering not only appearance, but also surrounding environment and temporal relations. In this paper, we propose a Symmetric Network with Spatial Relationship Modeling (SSM) method for retrieval. Specifically, design symmetric network learn unified cross-modal representations between descriptions images, where appearance details...

Super-resolution imaging based on single molecule localization of cellular structures nanometer scale requires to record a series wide-field or TIRF images resulting in considerable recording time (typically minutes). Therefore, sample drift becomes critical problem and will lower the precision. Herein we utilized morphological features specimen (mammalian cells) itself as reference markers replacing traditionally used (e.g., artificial fiduciary markers, fluorescent beads, metal...

Aspect sentiment quad prediction (ASQP) aims to predict the elements for a given sentence, which is critical task in field of aspect-based analysis. However, data imbalance issue has not received sufficient attention ASQP task. In this paper, we divide into two-folds, quad-pattern and aspect-category imbalance, propose an Adaptive Data Augmentation (ADA) framework tackle issue. Specifically, augmentation process with condition function adaptively enhances tail patterns aspect categories,...

Chinese Spelling Correction (CSC) aims to detect and correct the misspelled characters in texts. Recent studies have achieved great success by incorporating phonetic information for task predictions. Still, existing methods suffer from two limitations: 1) The differentiated between textual Pinyin pronunciation are underexplored. 2) predictions performed with over-emphasises on either or sequence, ignoring balanced modeling adaptive fusion. In this work, we proposed a method alleviate issues...

Multimodal Entity Linking (MEL) aims to link ambiguous mentions in multimodal contexts entities a knowledge graph. A pivotal challenge is fully leverage multi-element correlations between and bridge modality gap enable fine-grained semantic matching. Existing methods attempt several local correlative mechanisms, relying heavily on the automatically learned attention weights, which may over-concentrate partial correlations. To mitigate this issue, we formulate correlation assignment problem...

Low-Rank Adaptation (LoRA) drives research to align its performance with full fine-tuning. However, significant challenges remain: (1) Simply increasing the rank size of LoRA does not effectively capture high-rank information, which leads a bottleneck.(2) MoE-style methods substantially increase parameters and inference latency, contradicting goals efficient fine-tuning ease application. To address these challenges, we introduce Mixture Ranks (MoR), learns rank-specific information for...

Large language model (LLM) role-playing has gained widespread attention, where the authentic character knowledge is crucial for constructing realistic LLM agents. However, existing works usually overlook exploration of LLMs' ability to detect characters' known errors (KKE) and unknown (UKE) while playing roles, which would lead low-quality automatic construction trainable corpus. In this paper, we propose a probing dataset evaluate in KKE UKE. The results indicate that even latest LLMs...