Minghao Wu

Minghao Wu

Auto-generated profile, based on publicly available information

Minghao Wu

Tencent AI Lab, Shenzhen, Guangdong, China

Research Focus & Metrics

Academic Career Length

13Years (2014-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
28 October 2024
Verified
Conference Paper
Open access
ZWZhanyu Wang
Zhanyu Wang
LWLongyue Wang
Longyue Wang
ZZZhen Zhao
Zhen Zhao
MWMinghao Wu
Minghao Wu
CLChenyang Lyu
Chenyang Lyu
ZTZhaopeng Tu
Zhaopeng Tu
Recent advances in Multimodal Large Language Models (MLLMs) have constituted a significant leap forward in the field, particularly in the processing of videos, which encompasses inherent challenges such as spatiotemporal relationships. However, existing MLLMs are predominantly focused on the comprehension of video inputs, with limited capabilities in generating video content. In this paper, we present GPT4Video, a unified framework that seamlessly and lightly integrates with LLMs, visual feature extractors, and stable diffusion generative models for cohesive video understanding and generation. Moreover, we explore a text-only finetuning approach to equip models for instruction-following and safeguarding in multimodal conversations, enhancing training efficiency and generalization capabilities. Additionally, we construct multi-turn and caption-interleaved datasets for finetuning and benchmarking MLLMs, which serve as solid resources for advancing this field. Through quantitative and qualitative assessments, GPT4Video demonstrates the following advantages: 1) The framework incorporates video generation ability without adding extra training parameters, ensuring seamless compatibility with various video generators. 2) The model achieves superior performances across a variety of benchmarks. For instance, it outperforms Valley by 11.8% on video question answering, and surpasses NExt-GPT by 2.3% on text-to-video generation. 3) As safety pioneers in open-source MLLMs, we developed finetuning and evaluation datasets, securing an F1 score exceeding 80% in blocking harmful content during understanding and generating videos. In general, GPT4Video shows potential to function as a real-life assistant, marked by its effectiveness, adaptability, and safety.

10 March 2014 in Applied Physics Letters
Verified
Journal Article
Closed access
PDF file
YZYe Zhang
Ye Zhang
HWHuaqiang Wu
Huaqiang Wu
MWMinghao Wu
Minghao Wu
NDNing Deng
Ning Deng
ZYZhiping Yu
Zhiping Yu
HQHe Qian
He Qian
In this Letter, the origins of current fluctuations of Al/AlOx/WOy/W bilayer resistive random access memory (RRAM) devices are investigated through detailed noise analysis. Random telegraph noise (RTN) measurements were performed on RRAMs with three different resistance states. An obvious RTN signal with 40.7% amplitude difference was found at high resistance state, and the trapping/de-trapping process leading to the RTN signal was studied in detail by extracting the trap energy from energy diagram. For median and low resistance states, the resistance fluctuations were 34.0% and 0.3%, respectively. To further study the RTN characteristics, the normalized power spectral density (PSD) was analyzed. It is found that, for one dominant-trap caused RTN phenomena, the normalized noise PSD behaves as 1/f 2 on the high resistance state; while for median and low resistance states, the noise follows 1/f rule, suggesting that the current fluctuations are associated with the envelop of multiple RTNs caused by traps located near/in the conductive filament. Based on the noise analyses in time and frequency domains, a conduction mechanism is proposed to describe the trap effects on the current fluctuations of different resistance states.

8 October 2024
Verified
Conference Paper
Open access
CZChiyu Zhang
Chiyu Zhang
YSYifei Sun
Yifei Sun
MWMinghao Wu
Minghao Wu
JCJun Chen
Jun Chen
JLJie Lei
Jie Lei
BLBo Long
Bo Long
Content-based recommendation systems play a crucial role in delivering personalized content to users in the digital world. In this work, we introduce EmbSum, a novel framework that enables offline pre-computations of users and candidate items while capturing the interactions within the user engagement history. By utilizing the pretrained encoder-decoder model and poly-attention layers, EmbSum derives User Poly-Embedding (UPE) and Content Poly-Embedding (CPE) to calculate relevance scores between users and candidate items. EmbSum actively learns the long user engagement histories by generating user-interest summary with supervision from large language model (LLM). The effectiveness of EmbSum is validated on two datasets from different domains, surpassing state-of-the-art (SoTA) methods with higher accuracy and fewer parameters. Additionally, the model’s ability to generate summaries of user interests serves as a valuable by-product, enhancing its usefulness for personalized content recommendations.