Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks46 from public data
- 52
- 47
- 34
- Evaluating Parameter-Efficient Transfer Learning Approaches on SURE Benchmark for Speech Understanding25
The Speech UndeRstanding Evaluation (SURE) benchmark for parameter-efficient learning for various speech processing tasks is introduced and a new adapter, ConvAdapter, based on 1D convolution is introduced, showing that ConvAdapter outperforms the standard adapters while showing comparable performance against prefix tuning and Low-Rank Adaptation.
- Noninvasive prenatal diagnosis for Duchenne muscular dystrophy based on the direct haplotype phasing22
Direct haplotyping of NIPD based on linked-read sequencing for DMD is accurate and its feasibility for clinical application in the case of Duchenne Muscular Dystrophy is feasibility.
- Sea Buckthorn Polyphenols Alleviate High-Fat-Diet-Induced Metabolic Disorders in Mice via Reprograming Hepatic Lipid Homeostasis Owing to Directly Targeting Fatty Acid Synthase13
It is revealed that SBP as a functional food alleviates HFD-induced metabolic disorders in mice via reprograming hepatic lipid homeostasis caused by targeting FAS, owing to four polyphenols directly interacting with FAS and cinaroside binding to FAS with good affinity.
- Certainty-Driven Consistency Loss for Semi-supervised Learning13
A novel certainty-driven consistency loss (CCL) to dynamically select data samples that have relatively low uncertainty to guide the student learn more meaningful and certain/reliable targets, and hence improve the quality of the gradients backpropagated to the student.
- The application of late amniocentesis: a retrospective study in a tertiary fetal medicine center in China12
It is reasonable to recommend late amniocentesis as an effective and reliable method to detect late-onset fetal abnormalities, however, chromosomal microarray and whole-exome sequencing may result in uncertain results like variants of uncertain significance.
- 10
- Effects of Salvia miltiorrhiza active compounds on placenta-mediated pregnancy complications9
In this review, the literature is summarized and comprehensive insights on S. miltiorrhiza and its phytochemical constituents, pharmacological activities, and on PMPCs are provided, which would be valuable to explore promising drugs.
- 8
- 8
- 7
- 7
- HGMAN: Multi-Hop and Multi-Answer Question Answering Based on Heterogeneous Knowledge Graph (Student Abstract)7
A novel model that converts the ranking of candidate answers into individual predictions for each candidate, named HGMAN, which is capable of capturing more informative representations for relations assisted by the authors' heterogeneous graph, which consists of multiple entity nodes and relation nodes.
- Empowering Cartilage Restructuring with Biodegradable Magnesium Doped-Silicon Based-Nanoplatforms: Sustained Delivery and Enhanced Differentiation Potential6
The findings suggest the potential application of the developed biodegradable MMOS-Ica@HA nanoplatforms with acceptable drug loading capacity and controlled drug release in chondrodysplasia treatment, which indicates a promising approach for the treatment of chondrodysplasia.
- Identifying key m6A-methylated lncRNAs and genes associated with neural tube defects via integrative MeRIP and RNA sequencing analyses5
Investigation of differentially m6A-modified lncRNAs and genes involved in neural tube defect (NTD) development found decreased expression levels of Mir100hg, Gm19265, G m10544, Malat1, and genes, such as Zfp236, Erc2, and Hmg20a, in the NTD group may contribute to the development of NTD.
- An induced pluripotent stem cell line (GZHMCi003-A) derived from a fetus with exon 3 heterozygous deletion in RUNX2 gene causing cleidocranial dysplasia5
An induced pluripotent stem cell line GZHMCi003-A derived from umbilical cord blood mononuclear cells (UCBMCs) of a fetus with heterozygous deletion of the exon 3 in RUNX2 gene is generated to study the pathological mechanism and the treatment of CCD.
- HyperTTS: Parameter Efficient Adaptation in Text to Speech using Hypernetworks4
A small learnable network that generates parameters of the Adapter blocks, allowing us to condition Adapters on speaker representations and making them dynamic, and promising results on the dynamic adaptation of adapter parameters using hypernetworks open up new avenues for domain-generic multi-speaker TTS systems.
- 4
- 3
- 3
- Integrating Subgraph-Aware Relation and Direction Reasoning for Question Answering3
A novel neural model is proposed, Relation-updated Direction-guided Answer Selector (RDAS), which converts relations in each subgraph to additional nodes to learn structure information and utilizes direction information to enhance the reasoning ability.
- Leveraging Parameter-Efficient Transfer Learning for Multi-Lingual Text-to-Speech Adaptation2
This paper proposes to integrate parameter-efficient transfer learning (PETL) methods such as adapters and hypernetwork with TTS architecture for multilingual speech synthesis and shows that PETL methods able to achieve comparable or even better performance compared to full fine-tuning with only $\sim$2.5\% tunable parameters.
- Efficacy and safety of Chinese medicine combined with balloon dilatation vs. balloon dilatation alone for achalasia patients: a systematic review and meta-analysis2
Chinese medicine plus balloon dilatation had better effects after treatment than balloons alone for achalasia patients, and the risk of bias of included studies adhered to Cochrane criteria guidelines.
- 2
- From Knowledge to Inference: Formalizing Specialized Public Health Reasoning on GlobalHealthAtlas1
This work introduces GlobalHealthAt-las, a large scale multilingual dataset, and presents a domain aligned evaluator distilled from high confidence judgments of diverse LLMs to assess outputs along six dimensions: Accuracy, Reasoning, Completeness, Consensus Alignment, Terminology Norms, and In-sightfulness.
- FAD-MIL: a weakly supervised fracture detection model based on X-ray images1
Fractures are among the most common presentations to emergency and orthopedic services, yet radiograph interpretation remains inconsistent, particularly in resource-limited and off-hours settings, and FAD-MIL addresses these limitations through three design choices, including a global–local dual-stream architecture that captures both whole-image context and fine-grained tile features.
- 1
- 1
- Identification of Adrenomedullin-Induced S-Nitrosylated Proteins in JEG-3 Placental Cells1
Investigation of the JEG-3 placental cells provided evidences that ADM regulated the nitric oxide signaling pathway and modulated trophoblast invasion, and future studies are required to understand the roles of surface expression of S-nitrosylated ANX II on trophOBlast functions.
- Counterfactual-Augmented Data for Multi-Hop Knowledge Base Question Answering1
A model-agnostic Counterfactual Samples Synthesizing (CSS) training scheme that creates two augmented instances, one per augmentation method, to augment the training set.
- Automated quantification of coronal balance in spinal deformity: a safety-aware clinical workflow–
A robust HRNet-based automated system for quantifying Shoulder Height Difference, Pelvic Obliquity (PO), and C7 Plumb Line Offset (C7 offset) using a dataset of 847 full-spine radiographs is developed and validated, providing a robust foundation for large-scale clinical data auditing and future prospective studies on long-term patient outcomes.
- –
- Performance, Heterogeneity, and Methodological Quality of YOLO-Based Models for Fracture Detection: A Systematic Review and Meta-Analysis–
YOLO-based models showed promising overall technical performance for fracture detection, but the literature was highly heterogeneous, and the wide prediction interval indicates that performance may vary substantially across settings, and the pediatric subgroup finding remains preliminary.
- –
- Risk factors and prediction model for delayed bleeding after cold snare polypectomy: a retrospective study–
The risk factors were history of hypertension, hyperlipidemia, antithrombotics use, antiplatelet use, anticoagulant use, abdominal operation, sigmoid colon lesion, hematoma, cold snare defect protrusion, polyp size, wound size, the grade of wound bleeding, and morphology of Ip and these factors were incorporated into the prediction model for DB after CSP.
- Corrigendum: Identifying key m6A-methylated lncRNAs and genes associated with neural tube defects via integrative MeRIP and RNA sequencing analyses–
This research presents a novel probabilistic approach to estimating the number of mitochondria in the response of the immune system to oxidative stress.
- –
- Perinatal depression and anxiety among Chinese women during the outbreak of the COVID-19 Delta variant–
Pregnant women who were determined to be underweight before pregnancy, primiparous, above 35 years of age, employed full-time, of middle income status, and had comfortable living arrangements were at higher risk of developing depression and anxiety during a COVID-19 outbreak.
- An induced pluripotent stem cell line (GZHMCi004-A) derived from a fetus with heterozygous G380R mutation in FGFR3 gene causing achondroplasia–
An induced pluripotent stem cell line GZHMCi004-A derived from umbilical cord blood mononuclear cells of a fetus with heterozygous G380R mutation in FGFR3 gene is generated to study the pathological mechanism and the treatment of ACH.
- –
- –
- A Novel Salient Region Detection Method Based on Hierarchical Spatial Information–
A salient region detection method which mainly includes image abstraction, saliency calculation and integration, which has a good performance even though the object consists of complex patterns.
- –
- Privacy Preserving Mechanism for Multi-sensitive attributes Based on Time sequence–
Experimental results show that the time sequence introduced to several published data set problems, combined with thepublished data set Markov chain thought towards preserve data privacy, can effectively prevent the information loss of privacy, and enhance the security of the data release.
Publication data from OpenAlex, with missing venues and authors filled in from Crossref; citation counts are the higher of OpenAlex and Semantic Scholar, last synced 2026-10-12. One-sentence summaries under some papers are written by Semantic Scholar’s model. Citation counts may be lower than on Google Scholar, which indexes more sources.
Report an error
Wrong papers, two people merged into one, or a profile that should not be here? Tell us and we will fix or hide it.