检索结果-南通市图书馆

Newton design:designing CNNs with the family of Newton's methods

维普期刊数据库评论

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Science China(Information Sciences) 2023年第6期66卷 122-137页

作者： Zhengyang SHEN Yibo YANG Qi SHE Changhu WANG Jinwen MA Zhouchen LIN School of Mathematical Sciences Peking University bytedance ai lab JD Explore Academy Key laboratory of Machine Perception School of Intelligence Science and TechnologyPeking University Pazhou lab

Nowadays, convolutional neural networks(CNNs)have led the developments of machine ***, most CNN architectures are obtained by manual design, which is empirical, time-consuming, and non-transparent. In this paper, we aim at offering better insight into CNN models from the perspective of optimization theory. We propose a unified framework for understanding and designing CNN architectures with the family of Newton’s methods, which is referred to as Newton design. Specifically, we observe that the standard feedforward CNN model(PlainNet)solves an optimization problem via a kind of quasi-Newton method. Interestingly, residual network(ResNet)can also be derived if we use a more general quasi-Newton method to solve this problem. Based on the above observations, we solve this problem via a better method,the Newton-conjugate-gradient(Newton-CG)method, which inspires Newton-CGNet. In the network design,we translate binary-value terms in the optimization schemes to dropout layers, so dropout modules naturally appear in the derived CNN structures with specific locations, rather than being an empirical training *** experiments on image classification and text categorization tasks verify that Newton-CGNets perform very competitively. Particularly, Newton-CGNets surpass their counterparts ResNets by over 4% on CIFAR-10 and over 10% on CIFAR-100, respectively.

关键词： CNN dropout optimization method network design Newton's method

同方期刊数据库

Recent Progresses in Deep Learning Based Acoustic Models

在线全文

学校读者我要写书评

暂无评论

IEEE/CAA Journal of Automatica Sinica 2017年第3期4卷 396-409页

作者： Dong Yu Jinyu Li Tencent ai lab Tencent Microsoft ai and Research Microsoft

In this paper,we summarize recent progresses made in deep learning based acoustic models and the motivation and insights behind the surveyed *** first discuss models such as recurrent neural networks(RNNs) and convolutional neural networks(CNNs) that can effectively exploit variablelength contextual information,and their various combination with other *** then describe models that are optimized end-to-end and emphasize on feature representations learned jointly with the rest of the system,the connectionist temporal classification(CTC) criterion,and the attention-based sequenceto-sequence translation *** further illustrate robustness issues in speech recognition systems,and discuss acoustic model adaptation,speech enhancement and separation,and robust training *** also cover modeling techniques that lead to more efficient decoding and discuss possible future directions in acoustic model research.

关键词： Attention model convolutional neural network(CNN) connectionist temporal classification(CTC) deep learning(DL) long short-term memory(LSTM) permutation invariant training speech adaptation speech processing speech recognition speech separation

Dance2MIDI:Dance-driven multi-instrument music generation

在线全文

学校读者我要写书评

暂无评论

Computational Visual Media 2024年第4期10卷 791-802页

作者： Bo Han Yuheng Li Yixuan Shen Yi Ren Feilin Han College of Computer Science and Technology Zhejiang UniversityHangzhou 310058China National University of Singapore Singapore 119077Singapore Speech&Audio Team Bytedance AI LabSingapore 048583Singapore Department of Film and TV Technology Beijing Film AcademyBeijing 100088China

Dance-driven music generation aims to generate musical pieces conditioned on dance *** works focus on monophonic or raw audio generation,while the multi-instrument scenario is *** challenges associated with dancedriven multi-instrument music(MIDI)generation are twofold:(i)lack of a publicly available multi-instrument MIDI and video paired dataset and(ii)the weak correlation between music and *** tackle these challenges,we have built the first multi-instrument MIDI and dance paired dataset(D2MIDI).Based on this dataset,we introduce a multi-instrument MIDI generation framework(Dance2MIDI)conditioned on dance ***,to capture the relationship between dance and music,we employ a graph convolutional network to encode the dance *** allows us to extract features related to dance movement and dance ***,to generate a harmonious rhythm,we utilize a transformer model to decode the drum track sequence,leveraging a cross-attention ***,we model the task of generating the remaining tracks based on the drum track as a sequence understanding and completion task.A BERTlike model is employed to comprehend the context of the entire music piece through self-supervised *** evaluate the music generated by our framework trained on the D2MIDI dataset and demonstrate that our method achieves state-of-the-art performance.

关键词： video understanding music generation symbolic music cross-modal learning self-supervision

Few-shot Learning for Named Entity Recognition Based on BERT and Two-level Model Fusion

维普期刊数据库评论

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Data Intelligence 2021年第4期3卷 568-577页

作者： Yuan Gong Lu Mao Changliang Li ai lab King Soft CorporationBeijing 100086China

Currently,as a basic task of military document information extraction,Named Entity Recognition(NER)for military documents has received great *** 2020,China Conference on Knowledge Graph and Semantic Computing(CCKS)and System Engineering Research Institute of Academy of Military Sciences(AMS)issued the NER task for test evaluation,which requires the recognition of four types of entities including Test Elements(TE),Performance Indicators(PI),System Components(SC)and Task Scenarios(TS).Due to the particularity and confidentiality of the military field,only 400 items of annotated data are provided by the *** this paper,the task is regarded as a few-shot learning problem for NER,and a method based on BERT and two-level model fusion is ***,the proposed method is based on several basic models fine tuned by BERT on the training ***,a two-level fusion strategy applied to the prediction results of multiple basic models is proposed to alleviate the over-fitting ***,the labeling errors are eliminated by *** method achieves F1 score of 0.7203 on the test set of the evaluation task.

关键词： Few-shot learning Named entity recognition BERT Two-level model fusion

Secretome of polarized macrophages:potential for targeting inflammatory dynamics in spinal cord injury

在线全文

学校读者我要写书评

暂无评论

Neural Regeneration Research 2025年第11期20卷 3231-3232页

作者： Andreia Monteiro Susana Monteiro Nuno A.Silva Life and Health Sciences Research Institute(ICVS) School of MedicineUniversity of MinhoBragaPortugal ICVS/3B’s Associate lab PT Government Associated LabBraga/GuimaraesPortugal

Spinal cord injury(SCI)involves an initial traumatic phase,followed by secondary events such as ischemia,increased blood-spinal cord barrier permeability,ionic disruption,glutamate excitotoxicity,and metabolic alterations.A pe rsistent and exagge rated inflammato ry response within the spinal cord accompanies these events(Lima et al.,2022).The complexity and interplay of these mechanisms exacerbate the initial injury,leading to a degenerative process at the injury *** the initial trauma is unavoidable,the secondary injury begins within minutes and can last for months,creating an optimal window for therapeutic intervention.

关键词： alterations injury companies

Syntax-guided text generation via graph neural network

在线全文

学校读者我要写书评

暂无评论

Science China(Information Sciences) 2021年第5期64卷 67-76页

作者： Qipeng GUO Xipeng QIU Xiangyang XUE Zheng ZHANG Shanghai Key laboratory of Intelligent Information Processing School of Computer ScienceFudan University NYU Shanghai and AWS Shanghai ai lab

Text generation is a fundamental and important task in natural language processing. Most of the existing models generate text in a sequential manner and have difficulty modeling complex dependency structures. In this paper, we treat the text generation task as a graph generation problem exploiting both syntactic and word-ordering relationships. Leveraging the framework of the graph neural network, we propose the word graph model. During the process, the model builds a sentence incrementally and maintains syntactic integrity via a syntax-driven, top-down, breadth-first generation process. Experimental results on both synthetic and real text generation tasks show the efficacy of our approach.

关键词： text generation deep learning graph neural network dependency parsing

DocPedia: unleashing the power of large multimodal model in the frequency domain for versatile document understanding

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

Science China(Information Sciences) 2024年第12期67卷 65-78页

作者： Hao FENG Qi LIU Hao LIU Jingqun TANG Wengang ZHOU Houqiang LI Can HUANG Department of Electronic Engineering and Information Science University of Science and Technology of China bytedance Inc.

In this work, we present DocPedia, a novel large multimodal model(LMM) for versatile OCRfree document understanding, capable of parsing images up to 2560 × 2560 resolution. Unlike existing studies that either struggle with high-resolution documents or give up the large language model thus vision or language ability constrained, our DocPedia directly processes visual input in the frequency domain rather than the pixel space. The unique characteristic enables DocPedia to capture a greater amount of visual and textual information using a limited number of visual tokens. To consistently enhance both the perception and comprehension abilities of our DocPedia, we develop a dual-stage training strategy and enrich instructions/annotations of all training tasks covering multiple document types. Extensive quantitative and qualitative experiments are conducted on various publicly available benchmarks and the results confirm the mutual benefits of jointly learning perception and comprehension tasks. The results provide further evidence of the effectiveness and superior performance of our DocPedia over other methods.

关键词： document understanding large multimodal model OCR-free high-resolution frequency

同方期刊数据库

Past review,current progress,and challenges ahead on the cocktail party problem

在线全文

学校读者我要写书评

暂无评论

Frontiers of Information Technology & Electronic Engineering 2018年第1期19卷 40-63页

作者： Yan-min QIAN Chao WENG Xuan-kai CHANG Shuai WANG Dong YU Tencent ai lab Tencent Bellevue 98004 USA Department of Computer Science and Engineering Shanghai Jiao Tong University Shanghai 200240 China

The cocktail party problem,i.e.,tracing and recognizing the speech of a specific speaker when multiple speakers talk simultaneously,is one of the critical problems yet to be solved to enable the wide application of automatic speech recognition(ASR) *** this overview paper,we review the techniques proposed in the last two decades in attacking this *** focus our discussions on the speech separation problem given its central role in the cocktail party environment,and describe the conventional single-channel techniques such as computational auditory scene analysis(CASA),non-negative matrix factorization(NMF) and generative models,the conventional multi-channel techniques such as beamforming and multi-channel blind source separation,and the newly developed deep learning-based techniques,such as deep clustering(DPCL),the deep attractor network(DANet),and permutation invariant training(PIT).We also present techniques developed to improve ASR accuracy and speaker identification in the cocktail party *** argue effectively exploiting information in the microphone array,the acoustic training set,and the language itself using a more powerful *** optimization ob jective and techniques will be the approach to solving the cocktail party problem.

关键词： Cocktail party problem Computational auditory scene analysis Non-negative matrix factorization Permutation invariant training Multi-talker speech processing

ChemDFM-X: towards large multimodal model for chemistry

在线全文

学校读者我要写书评

暂无评论

Science China(Information Sciences) 2024年第12期67卷 99-100页

作者： Zihan ZHAO Bo CHEN Jingpiao LI Lu CHEN Liyang WEN Pengyu WANG Zichen ZHU Danyang ZHANG Yansi LI Zhongyang Dai Xin CHEN Kai YU X-LANCE lab Department of Computer Science and EngineeringMoE Key Lab of Artificial Intelligence AI Institute Shanghai Jiao Tong University Suzhou laboratory

Chemistry, as a naturally multimodal discipline, plays a crucial role in various vital fields such as pharmaceutical research and material manufacturing. Therefore, research on artificial intelligence(ai) for chemistry has garnered increasing attention. Despite the rapid development, most of the chemical ai models today mainly focus on single tasks with unimodal input [1].

关键词：

同方期刊数据库