30,372 Hits in 4.6 sec

Insight DCU at TRECVid 2015

Kevin McGuinness, Eva Mohedano, Amaia Salvador, Zhenxing Zhang, Mark Marsden, Peng Wang, Iveel Jargalsaikhan, Joseph Antony, Xavier Giró-i-Nieto, Shin'ichi Satoh, Noel E. O'Connor, Alan F. Smeaton
2015 TREC Video Retrieval Evaluation  
In the INS task we used deep convolutional network features trained on external data and the query data for this year to train our system.  ...  In the SIN task we again used convolutional network features, this time finetuning a network pretrained on external data for the task.  ...  We gratefully acknowledge the support of NVIDIA Corporation with the donation of the GPUs used for this research. Programme material for the instance search task is c BBC.  ... 
dblp:conf/trecvid/McGuinnessMSZM015 fatcat:nbjhhc3qozcvri3vmttbckgu7u

Table of Contents

2021 IEEE transactions on multimedia  
Fang Deep Learning for Multimedia Processing Progressive Learning of Low-Precision Networks for Image Classification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...Hierarchical  ...  Quan Multimedia Search and Retrieval Progressive Bilateral-Context Driven Model for Post-Processing Person Re-Identification . . . . . . . . . . . . . . . . . . . . . . . . . ...Dual Neural Networks Coupling  ... 
doi:10.1109/tmm.2021.3132246 fatcat:el7u2udtybddrpbl5gxkvfricy

DeepLight: Deep Lightweight Feature Interactions for Accelerating CTR Predictions in Ad Serving [article]

Wei Deng and Junwei Pan and Tian Zhou and Deguang Kong and Aaron Flores and Guang Lin
2021 arXiv   pre-print
The embedding-based neural networks have been proposed to learn both explicit feature interactions through a shallow component and deep feature interactions using a deep neural network (DNN) component.  ...  : 1) accelerate the model inference via explicitly searching informative feature interactions in the shallow component; 2) prune redundant layers and parameters at intra-layer and inter-layer level in  ...  feature interactions and a DNN * Equal contribution component for powerful non-linear modeling.  ... 
arXiv:2002.06987v3 fatcat:a4lvvhi7bnanlp2cd5sqqqownu

Deep Collaborative Filtering: A Recommendation Method for Crowdfunding Project Based on the Integration of Deep Neural Network and Collaborative Filtering

Pei Yin, Jing Wang, Jun Zhao, Huan Wang, Hongcheng Gan, Wei Liu
2022 Mathematical Problems in Engineering  
In the perspective of implicit feedback, this method uses the advantages of convolutional neural network for effective learning of the nonlinear interaction of users and items and the characteristics of  ...  collaborative filtering algorithm for modeling the linear interaction of users and items and combines the two methods for recommendation.  ...  Concatenate the latent feature vectors f 1 and f 2 learned by matrix factorization and deep neural network.  ... 
doi:10.1155/2022/4655030 fatcat:6nadmi32g5hrzaurn3bukqazzi

Curriculum Learning Based on Reward Sparseness for Deep Reinforcement Learning of Task Completion Dialogue Management

Atsushi Saito
2018 Proceedings of the 2018 EMNLP Workshop SCAI: The 2nd International Workshop on Search-Oriented Conversational AI  
We also propose a dialogue policy based on progressive neural networks whose modules with parameters are appended with previous parameters fixed as the curriculum proceeds, and this policy improves performances  ...  This curriculum makes it possible to learn dialogue management for sets of user goals with large number of slots.  ...  Acknowledgements The author wishes to thank Taichi Iki, Yuki Sekizawa, and the anonymous reviewers for helpful comments.  ... 
doi:10.18653/v1/w18-5707 dblp:conf/emnlp/Saito18 fatcat:7xxn6g7ssrdz3lwsuynsuheqja

Automated Machine Learning for Deep Recommender Systems: A Survey [article]

Bo Chen, Xiangyu Zhao, Yejing Wang, Wenqi Fan, Huifeng Guo, Ruiming Tang
2022 arXiv   pre-print
Then we discuss the state-of-the-art AutoML approaches that automate the feature selection, feature embeddings, feature interactions, and system design in DRS.  ...  Despite their advancements, DRS models, like other deep learning models, employ sophisticated neural network architectures and other vital components that are typically designed and tuned by human experts  ...  : Typically, the features for DRS are high-dimensional and extremely sparse.  ... 
arXiv:2204.01390v1 fatcat:ybiang7gajdkrljsrhbq6ih62m

NASRec: Weight Sharing Neural Architecture Search for Recommender Systems [article]

Tunhou Zhang, Dehua Cheng, Yuchen He, Zhengxing Chen, Xiaoliang Dai, Liang Xiong, Feng Yan, Hai Li, Yiran Chen, Wei Wen
2022 arXiv   pre-print
The rise of deep neural networks provides an important driver in optimizing recommender systems.  ...  However, the success of recommender systems lies in delicate architecture fabrication, and thus calls for Neural Architecture Search (NAS) to further improve its modeling.  ...  We also thank Maxim Naumov, Jeff Hwang and Colin Taylor in Meta Platforms, Inc. for their kind help on this project. 5  ... 
arXiv:2207.07187v1 fatcat:2mhzriarzjc6vffbticztpz5ti

Visual Design of Artificial Intelligence Based on the Image Search Algorithm

Xiaobo Jiang
2020 Journal of Applied Data Sciences  
Experiments show that the AI visualization design based on the image search algorithm can not only overcome the semantic gap to some extent, but also strengthen the interaction between 88% systems and  ...  users to browse the search results more efficiently and naturally.  ...  with various other unsupervised and supervised sparse coding and supervised convolutional methods for convolutional neural networks on MINST.  ... 
doi:10.47738/jads.v1i2.56 fatcat:fwkvlhiymjhqnblckpwkxxydtm

Qualitative similarities and differences in visual object representations between brains and deep networks

Georgin Jacob, R. T. Pramod, Harish Katti, S. P. Arun
2021 Nature Communications  
These findings indicate sufficient conditions for the emergence of these phenomena in brains and deep networks, and offer clues to the properties that could be incorporated to improve deep networks.  ...  Here, we recast well-known perceptual and neural phenomena in terms of distance comparisons, and ask whether they are present in feedforward deep neural networks trained for object recognition.  ...  To characterize the progression of the scene incongruence effect, we calculated the distance between each scene (object + context) to the average feature vector for the object.  ... 
doi:10.1038/s41467-021-22078-3 pmid:33767141 fatcat:iba4wgs5sfei3mhgmr4gq4x7k4

Editorial for the ICMR 2017 special issue

Michael S. Lew
2018 International Journal of Multimedia Information Retrieval  
The work "Balancing Search Space Partitions by Sparse Coding for Distributed Redundant Media Indexing and Retrieval" by Andre Mourao and Joao Magalhaes addresses the challenges on a modern distributed  ...  Conclusions are given on which features are most effective in classifying search intent.  ... 
doi:10.1007/s13735-018-0148-0 fatcat:nbgvh2lgmfhixm6zcl4ig455ja

A Review on Automated Disease Diagnosis Techniques

Sunena Rose M V, Dr. Sobhana N. V
2017 IJARCCE  
In this paper, a survey of different techniques for automatic disease diagnosis is done.  ...  The development of computer technologies and increased expenditure of healthcare are the reasons for innovation of automated disease inference system.  ...  [13] presented a sparse deep learning algorithm for recognition and categorization.  ... 
doi:10.17148/ijarcce.2017.63187 fatcat:bxgqlbcyyzc5tjstmb6xdc6ogi

Alzheimer's Disease Diagnosis via Deep Factorization Machine Models [article]

Raphael Ronge and Kwangsik Nho and Christian Wachinger and Sebastian Pölsterl
2021 arXiv   pre-print
The current state-of-the-art deep neural networks (DNNs) for Alzheimer's Disease diagnosis use different biomarker combinations to classify patients, but do not allow extracting knowledge about the interactions  ...  The proposed model has three parts: (i) an embedding layer to deal with sparse categorical data, (ii) a Factorization Machine to efficiently learn pairwise interactions, and (iii) a DNN to implicitly model  ...  The third part is a Deep Neural Network (DNN) that has the potential to implicitly learn complex feature interactions.  ... 
arXiv:2108.05916v1 fatcat:jpzlc6as2vf6jjvru6rt5f7tym

Field-aware Neural Factorization Machine for Click-Through Rate Prediction

Li Zhang, Weichen Shen, Jianhang Huang, Shijian Li, Gang Pan
2019 IEEE Access  
This paper can have strong second-order feature interactive learning ability, such as Field-aware Factorization Machine; on this basis, a deep neural network is used for higher order feature combination  ...  This paper combines the traditional feature combination methods and the deep neural networks to automate the feature combinations to improve the accuracy of the click-through rate prediction.  ...  For other models, search from 4, 8, 16, 32, 64 . The structure of the deep neural network is searched from a combination of 2 or 3 layers and 128 or 256 neurons per layer.  ... 
doi:10.1109/access.2019.2921026 fatcat:5kqan5ntq5hexmcoetguclbequ

A Modified Grasshopper Optimization Algorithm Combined with Convolutional Neural Network for Content Based Image Retrieval

2019 International Journal of Engineering  
In this paper, a convolutional neural network (CNN) is used to extract deep and high-level features from the images. Next, an optimization problem is defined in order to model the retrieval system.  ...  Although many efficient researches have been performed for this topic so far, there is a semantic gap between human concept and features extracted from the images and it has become an important problem  ...  Example of a convolutional neural network for classifying images [24] (1) where the position of the i-th grasshopper is social interaction is represented by and shows the wind influence.  ... 
doi:10.5829/ije.2019.32.07a.04 fatcat:seqfwdjy2naf5j3ptrp6xvacye

Neural best-buddies

Kfir Aberman, Jing Liao, Mingyi Shi, Dani Lischinski, Baoquan Chen, Daniel Cohen-Or
2018 ACM Transactions on Graphics  
This paper presents a novel method for sparse cross-domain correspondence.  ...  Our approach operates on hierarchies of deep features, extracted from the input images by a pre-trained CNN.  ...  ACKNOWLEDGMENTS We thank the anonymous reviewers for their helpful comments.  ... 
doi:10.1145/3197517.3201332 fatcat:4gh4tjhjpzattcpnonyltco4hu
« Previous Showing results 1 — 15 out of 30,372 results