Search

Your search keyword '"Ma, Wentao"' showing total 130 results
Start Over "Ma, Wentao" Publication Year Range This year
130 results on '"Ma, Wentao"'

1. FlowBench: Revisiting and Benchmarking Workflow-Guided Planning for LLM-based Agents

Author

Xiao, Ruixuan, Ma, Wentao, Wang, Ke, Wu, Yuchuan, Zhao, Junbo, Wang, Haobo, Huang, Fei, and Li, Yongbin

Subjects
Computer Science - Computation and Language
Abstract

LLM-based agents have emerged as promising tools, which are crafted to fulfill complex tasks by iterative planning and action. However, these agents are susceptible to undesired planning hallucinations when lacking specific knowledge for expertise-intensive tasks. To address this, preliminary attempts are made to enhance planning reliability by incorporating external workflow-related knowledge. Despite the promise, such infused knowledge is mostly disorganized and diverse in formats, lacking rigorous formalization and comprehensive comparisons. Motivated by this, we formalize different formats of workflow knowledge and present FlowBench, the first benchmark for workflow-guided planning. FlowBench covers 51 different scenarios from 6 domains, with knowledge presented in diverse formats. To assess different LLMs on FlowBench, we design a multi-tiered evaluation framework. We evaluate the efficacy of workflow knowledge across multiple formats, and the results indicate that current LLM agents need considerable improvements for satisfactory planning. We hope that our challenging benchmark can pave the way for future agent planning research.

Published
2024

2. PTA: Enhancing Multimodal Sentiment Analysis through Pipelined Prediction and Translation-based Alignment

Author

Song, Shezheng, Li, Shasha, Zhao, Shan, Wang, Chengyu, Li, Xiaopeng, Yu, Jie, Wan, Qian, Ma, Jun, Yan, Tianwei, Ma, Wentao, and Mao, Xiaoguang

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

Multimodal aspect-based sentiment analysis (MABSA) aims to understand opinions in a granular manner, advancing human-computer interaction and other fields. Traditionally, MABSA methods use a joint prediction approach to identify aspects and sentiments simultaneously. However, we argue that joint models are not always superior. Our analysis shows that joint models struggle to align relevant text tokens with image patches, leading to misalignment and ineffective image utilization. In contrast, a pipeline framework first identifies aspects through MATE (Multimodal Aspect Term Extraction) and then aligns these aspects with image patches for sentiment classification (MASC: Multimodal Aspect-Oriented Sentiment Classification). This method is better suited for multimodal scenarios where effective image use is crucial. We present three key observations: (a) MATE and MASC have different feature requirements, with MATE focusing on token-level features and MASC on sequence-level features; (b) the aspect identified by MATE is crucial for effective image utilization; and (c) images play a trivial role in previous MABSA methods due to high noise. Based on these observations, we propose a pipeline framework that first predicts the aspect and then uses translation-based alignment (TBA) to enhance multimodal semantic consistency for better image utilization. Our method achieves state-of-the-art (SOTA) performance on widely used MABSA datasets Twitter-15 and Twitter-17. This demonstrates the effectiveness of the pipeline approach and its potential to provide valuable insights for future MABSA research. For reproducibility, the code and checkpoint will be released., Comment: Code will be released upon publication

Published
2024

5. Exosome is a Fancy Mobile Sower of Ferroptosis

Author

Hou, Qin, Ouyang, Siyu, Xie, Zhongcheng, He, Yinling, Deng, Yunong, Guo, Jiamin, Yu, Panpan, Tan, Xiaoqian, Ma, Wentao, Li, Pin, Yu, Jiang, Mo, Qinger, Zhang, Zhixia, Chen, Dandan, Lin, Xiaoyan, Liu, Zhiyang, Chen, Xi, Peng, Tianhong, Li, Liang, and Xie, Wei

Published
2024
Full Text
View/download PDF

42. Study on Foaming Agent Foam Composite Index (FCI) Correlation with High Temperature and High Pressure for Unconventional Oil and Gas Reservoirs.

Author

Wu, Jianjun, Ma, Wentao, Liu, Yinhua, Qi, Wei, Wang, Haoyu, Ji, Guofa, Luo, Wei, and Liu, Kai

Subjects
PETROLEUM reservoirs, GAS wells, DRILLING fluids, DRILLING muds, HIGH temperatures, FOAM, SURFACE active agents, GAS reservoirs
Abstract

In the process of unconventional oil and gas reservoir exploitation, it is difficult to reduce drilling fluid lost in natural fractures, enhance the CO2 displacement effect and reduce foam drainage gas recovery costs. In most cases, foaming agents can solve these problems in a low-cost way in a short period of time. Foaming agent screening and evaluation is the key to this technology. However, there are few experimental tests used in the evaluation of foaming agent properties that match the actual unconventional oil or gas well conditions of high temperature and high pressure. Using the actual temperature and pressure conditions of a wellbore, the foaming capacity and half-life of two common foaming agents were systematically evaluated by using the high-temperature and high-pressure visual foam properties evaluation device (UPMX-500), in which the foaming agent's volume concentration was 3‰ in a simulated formation water with a pH of 6 and salinity of 9 × 104 mg/L. The high-temperature (40 °C, 60 °C, 80 °C, 100 °C) and high-pressure (0.1 MPa, 6.0 MPa, 8.0 MPa, 10.0 MPa) effect on the foaming capacity and half-life was analyzed. Binary linear regression of pressure and temperature was carried out, taking the foam composite index as the target and using a formula with high correlation. The results showed that the foam composite index (FCI) of the two foaming agents was positively correlated with pressure and temperature. The correlation of UT-7 was FCI = 64.1196T + 735.713p − 2066.2, the correlation of HY-3K was FCI = 62.5523T + 7220.391p − 2415.6, and the coefficients of determination were 0.9799 and 0.9895, respectively, with an error of less than 10%. This correlation equation can provide a reference for accurately predicting the foaming capacity of foaming agents under high-temperature and high-pressure conditions and can also be used to optimize foaming agents or to qualitatively evaluate results for the efficient exploitation of unconventional oil and gas reservoirs. [ABSTRACT FROM AUTHOR]

Published
2024
Full Text
View/download PDF

Catalog

Books, media, physical & digital resources
See catalog results