Search

Your search keyword '"Cheng, Weiwei"' showing total 1,022 results
Start Over Author "Cheng, Weiwei"
1,022 results on '"Cheng, Weiwei"'

1. Retrieve, Annotate, Evaluate, Repeat: Leveraging Multimodal LLMs for Large-Scale Product Retrieval Evaluation

Author

Hosseini, Kasra, Kober, Thomas, Krapac, Josip, Vollgraf, Roland, Cheng, Weiwei, and Ramallo, Ana Peleteiro

Subjects
Computer Science - Information Retrieval, Computer Science - Artificial Intelligence, Computer Science - Computation and Language, Computer Science - Emerging Technologies, Computer Science - Human-Computer Interaction
Abstract

Evaluating production-level retrieval systems at scale is a crucial yet challenging task due to the limited availability of a large pool of well-trained human annotators. Large Language Models (LLMs) have the potential to address this scaling issue and offer a viable alternative to humans for the bulk of annotation tasks. In this paper, we propose a framework for assessing the product search engines in a large-scale e-commerce setting, leveraging Multimodal LLMs for (i) generating tailored annotation guidelines for individual queries, and (ii) conducting the subsequent annotation task. Our method, validated through deployment on a large e-commerce platform, demonstrates comparable quality to human annotations, significantly reduces time and cost, facilitates rapid problem discovery, and provides an effective solution for production-level quality control at scale., Comment: 13 pages, 5 figures, 4 Tables

Published
2024

2. What should I wear to a party in a Greek taverna? Evaluation for Conversational Agents in the Fashion Domain

Author

Maronikolakis, Antonis, Ramallo, Ana Peleteiro, Cheng, Weiwei, and Kober, Thomas

Subjects
Computer Science - Information Retrieval, Computer Science - Artificial Intelligence, Computer Science - Computation and Language, Computer Science - Human-Computer Interaction
Abstract

Large language models (LLMs) are poised to revolutionize the domain of online fashion retail, enhancing customer experience and discovery of fashion online. LLM-powered conversational agents introduce a new way of discovery by directly interacting with customers, enabling them to express in their own ways, refine their needs, obtain fashion and shopping advice that is relevant to their taste and intent. For many tasks in e-commerce, such as finding a specific product, conversational agents need to convert their interactions with a customer to a specific call to different backend systems, e.g., a search system to showcase a relevant set of products. Therefore, evaluating the capabilities of LLMs to perform those tasks related to calling other services is vital. However, those evaluations are generally complex, due to the lack of relevant and high quality datasets, and do not align seamlessly with business needs, amongst others. To this end, we created a multilingual evaluation dataset of 4k conversations between customers and a fashion assistant in a large e-commerce fashion platform to measure the capabilities of LLMs to serve as an assistant between customers and a backend engine. We evaluate a range of models, showcasing how our dataset scales to business needs and facilitates iterative development of tools., Comment: Accepted at KDD workshop on Evaluation and Trustworthiness of Generative AI Models

Published
2024

19. Grounding Natural Language Instructions: Can Large Language Models Capture Spatial Information?

Author

Rozanova, Julia, Ferreira, Deborah, Dubba, Krishna, Cheng, Weiwei, Zhang, Dell, and Freitas, Andre

Subjects
Computer Science - Computation and Language, Computer Science - Artificial Intelligence
Abstract

Models designed for intelligent process automation are required to be capable of grounding user interface elements. This task of interface element grounding is centred on linking instructions in natural language to their target referents. Even though BERT and similar pre-trained language models have excelled in several NLP tasks, their use has not been widely explored for the UI grounding domain. This work concentrates on testing and probing the grounding abilities of three different transformer-based models: BERT, RoBERTa and LayoutLM. Our primary focus is on these models' spatial reasoning skills, given their importance in this domain. We observe that LayoutLM has a promising advantage for applications in this domain, even though it was created for a different original purpose (representing scanned documents): the learned spatial features appear to be transferable to the UI grounding setting, especially as they demonstrate the ability to discriminate between target directions in natural language instructions., Comment: *Equal contribution

Published
2021

38. Metformin ameliorates the mitochondrial damage induced by C9ORF72 amyotrophic lateral sclerosis/frontotemporal dementia-related poly-GR

Author

FENG Yiyuan, XU Zhongyun, YIN Yafu, WANG Hui, and CHENG Weiwei

Subjects
c9orf72 amyotrophic lateral sclerosis/ frontotemporal dementia (c9orf72 als/ftd), poly-glycine-arginine (poly-gr), mitochondrion, phosphorylated protein kinase b, metformin, Medicine
Abstract

Objective·To investigate the effect of C9ORF72 amyotrophic lateral sclerosis/frontotemporal dementia (ALS/FTD)-related poly-glycine-arginine (poly-GR) on mitochondrial morphology and function, and analyze the rescue effect of metformin on mitochondrial damage induced by poly-GR and its underlying mechanism.Methods·SK-N-SH cells stably overexpressing 50 repeated glycine-arginine sequences [(GR)50] or green fluorescent protein (GFP) were constructed by lentivirus infection, which were respectively named as (GR)50-SK cell line and GFP CTRL-SK cell line. (GR)50 expression in (GR)50-SK cells was verified by Western blotting. GFP expression in GFP GTRL-SK cells was observed by fluorescence microscope. Propidium iodide (PI) staining was used to detect the apoptosis levels of (GR)50-SK and GFP CTRL-SK cells. Immunofluorescence (IF) staining was performed to determine the subcellular location of (GR)50. Reactive oxygen species (ROS) level of mitochondria was evaluated by staining cells with MitoSOX Red followed by observing the intensity of red fluorescence under fluorescence microscope. The mitochondrial morphology of (GR)50-SK and GFP CTRL-SK cells was observed by transmission electron microscopy. Western blotting was used to detect protein kinase B (PKB, also known as AKT) and its phosphorylation levels in (GR)50-SK and GFP CTRL-SK cells. SC79 was used to activate AKT in (GR)50-SK cells, and MitoSOX Red staining and PI staining were used to analyze mitochondrial ROS and apoptosis levels after phosphorylated AKT increased. Metformin was used to treat (GR)50-SK and GFP CTRL-SK cells, respectively, and the apoptosis levels, mitochondrial ROS levels, mitochondrial morphology, AKT and its phosphorylation levels, and ATP concentrations of the two cells were detected by the above methods and ATP detection kit, respectively.Results·Western blotting showed that the construction of (GR)50-SK cells was successful, and fluorescence microscopy showed that the construction of GFP CTRL-SK cells was also successful. PI staining results showed that the apoptosis level of (GR)50-SK cells was higher than that of the GFP CTRL-SK cells (P=0.016). IF staining results showed that there was partial co-localization of (GR)50 in the mitochondria of (GR)50-SK cells. Compared with GFP CTRL-SK cells, the mitochondrial morphology and structure of (GR)50-SK cells were significant abnormalities, with a significantly increased ROS levels. The AKT levels in (GR)50-SK cells were similar to those in the GFP CTRL-SK cells, but there was a significant decrease in phosphorylated AKT levels. After (GR)50-SK cells were treated with SC79, the AKT phosphorylation level was significantly upregulated, and ROS level and apoptosis level were significantly downregulated. Metformin could significantly up-regulate the phosphorylated AKT levels in (GR)50-SK cells, but had no effect on AKT levels; it could reshape the morphology and structure of some mitochondria, reduce ROS levels, increase ATP production (P=0.000), and down-regulate the level of cell apoptosis (P=0.000).Conclusion·(GR)50 can cause mitochondrial morphology and function abnormalities by down-regulating AKT phosphorylation, and promote cell apoptosis. Metformin can effectively reduce the occurrence of the above pathological events induced by (GR)50.

Published
2023
Full Text
View/download PDF

39. Research progress of integrated stress response in pathogenesis of Alzheimer's disease

Author

SUN Hui, JIN Hongfu, GUO Shenrui, FENG Yiyuan, YIN Yafu, WANG Hui, and CHENG Weiwei

Subjects
integrated stress response, alzheimer disease, integrated stress response inhibitor, protein homeostasis, Medicine
Abstract

Integrated stress response (ISR) is a cellular adaptive response induced by stress, which is strictly regulated by multiple phosphokinases, phosphatases and other proteins to maintain protein homeostasis. Studies have shown that ISR is abnormally activated in Alzheimer's disease, and targeted regulation of different proteins in ISR pathway inhibits the abnormal activation of ISR, leading to restoration of protein homeostasis and alleviation of the neuropathological changes and memory impairment in Alzheimer's disease models. These lines of evidence suggest that ISR has the potential to be a therapeutic target in Alzheimer's disease treatment. This paper reviews the abnormal activation and regulation mechanism of ISR in Alzheimer's disease and discusses the application of ISR as therapeutic targets to Alzheimer's disease models.

Published
2023
Full Text
View/download PDF

40. Evaluating for Diversity in Question Generation over Text

Author

Schlichtkrull, Michael Sejr and Cheng, Weiwei

Subjects
Computer Science - Computation and Language, Statistics - Machine Learning
Abstract

Generating diverse and relevant questions over text is a task with widespread applications. We argue that commonly-used evaluation metrics such as BLEU and METEOR are not suitable for this task due to the inherent diversity of reference questions, and propose a scheme for extending conventional metrics to reflect diversity. We furthermore propose a variational encoder-decoder model for this task. We show through automatic and human evaluation that our variational model improves diversity without loss of quality, and demonstrate how our evaluation scheme reflects this improvement.

Published
2020

49. CRISPR-Cas9 Screens Identify the RNA Helicase DDX3X as a Repressor of C9ORF72 (GGGGCC)n Repeat-Associated Non-AUG Translation

Author

Cheng, Weiwei, Wang, Shaopeng, Zhang, Zhe, Morgens, David W, Hayes, Lindsey R, Lee, Soojin, Portz, Bede, Xie, Yongzhi, Nguyen, Baotram V, Haney, Michael S, Yan, Shirui, Dong, Daoyuan, Coyne, Alyssa N, Yang, Junhua, Xian, Fengfan, Cleveland, Don W, Qiu, Zhaozhu, Rothstein, Jeffrey D, Shorter, James, Gao, Fen-Biao, Bassik, Michael C, and Sun, Shuying

Subjects
Biological Psychology, Biomedical and Clinical Sciences, Neurosciences, Psychology, Frontotemporal Dementia (FTD), Brain Disorders, Neurodegenerative, ALS, Acquired Cognitive Impairment, Dementia, Rare Diseases, Genetics, 2.1 Biological and endogenous factors, Aetiology, Neurological, Amyotrophic Lateral Sclerosis, Animals, C9orf72 Protein, CRISPR-Cas Systems, DEAD-box RNA Helicases, Drosophila, Frontotemporal Dementia, Humans, Protein Biosynthesis, Repetitive Sequences, Nucleic Acid, CRISPR-Cas9 screen, DDX3X, FTD, RAN translation, RNA, helicase, neurodegeneration, repeat expansion, Cognitive Sciences, Neurology & Neurosurgery, Biological psychology
Abstract

Hexanucleotide GGGGCC repeat expansion in C9ORF72 is the most prevalent genetic cause of amyotrophic lateral sclerosis (ALS) and frontotemporal dementia (FTD). One pathogenic mechanism is the aberrant accumulation of dipeptide repeat (DPR) proteins produced by the unconventional translation of expanded RNA repeats. Here, we performed genome-wide CRISPR-Cas9 screens for modifiers of DPR protein production in human cells. We found that DDX3X, an RNA helicase, suppresses the repeat-associated non-AUG translation of GGGGCC repeats. DDX3X directly binds to (GGGGCC)n RNAs but not antisense (CCCCGG)n RNAs. Its helicase activity is essential for the translation repression. Reduction of DDX3X increases DPR levels in C9ORF72-ALS/FTD patient cells and enhances (GGGGCC)n-mediated toxicity in Drosophila. Elevating DDX3X expression is sufficient to decrease DPR levels, rescue nucleocytoplasmic transport abnormalities, and improve survival of patient iPSC-differentiated neurons. This work identifies genetic modifiers of DPR protein production and provides potential therapeutic targets for C9ORF72-ALS/FTD.

Published
2019

Catalog

Books, media, physical & digital resources
See catalog results