Filters








718 Hits in 6.1 sec

DCN V2: Improved Deep Cross Network and Practical Lessons for Web-scale Learning to Rank Systems [article]

Ruoxi Wang, Rakesh Shivanna, Derek Z. Cheng, Sagar Jain, Dong Lin, Lichan Hong, Ed H. Chi
2020 arXiv   pre-print
Unfortunately, in models that serve web-scale traffic with billions of training examples, DCN showed limited expressiveness in its cross network at learning more predictive feature interactions.  ...  Despite significant research progress made, many deep learning models in production still rely on traditional feed-forward neural networks to learn feature crosses inefficiently.  ...  Unfortunately, this involves a combinatorial search space, which is large and sparse in web-scale applications where the data is mostly categorical.  ... 
arXiv:2008.13535v2 fatcat:eidmwxuburbirh5jbxgnyqcsyu

Deep Learning in Science [article]

Stefano Bianchini, Moritz Müller, Pierre Pelletier
2020 arXiv   pre-print
Much of the recent success of Artificial Intelligence (AI) has been spurred on by impressive achievements within a broader family of machine learning methods, commonly referred to as Deep Learning (DL)  ...  Therefore, we empirically assess how DL adoption relates to re-combinatorial novelty and scientific impact in the health sciences.  ...  Put simply, without neural networks, there would be no deep learning.  ... 
arXiv:2009.01575v2 fatcat:4ttqgjdjfjbydp7flnhcgg5p7m

Deep Learning in the Wild [chapter]

Thilo Stadelmann, Mohammadreza Amirian, Ismail Arabaci, Marek Arnold, Gilbert François Duivesteijn, Ismail Elezi, Melanie Geiger, Stefan Lörwald, Benjamin Bruno Meier, Katharina Rombach, Lukas Tuggener
2018 Lecture Notes in Computer Science  
While this interest is fueled by beautiful success stories, practical work in deep learning on novel tasks without existing baselines remains challenging.  ...  providing best practices for deep learning in practice.  ...  A next step will explore an explicitly hierarchical learner to cope with the combinatorial explosion of the action space on the three time scales (operational/tactical/strategic) without using hard-coded  ... 
doi:10.1007/978-3-319-99978-4_2 fatcat:gmpuuzlio5ea3ck75fekk3ab7y

Deep Learning in the Wild [article]

Thilo Stadelmann, Mohammadreza Amirian and Ismail Arabaci, Marek Arnold, Gilbert François Duivesteijn, Ismail Elezi, Melanie Geiger and Stefan Lörwald and Benjamin Bruno Meier, Katharina Rombach, Lukas Tuggener
2018 arXiv   pre-print
While this interest is fueled by beautiful success stories, practical work in deep learning on novel tasks without existing baselines remains challenging.  ...  providing best practices for deep learning in practice.  ...  A next step will explore an explicitly hierarchical learner to cope with the combinatorial explosion of the action space on the three time scales (operational/tactical/strategic) without using hard-coded  ... 
arXiv:1807.04950v1 fatcat:6cb63xget5fynmjxrhzcpirvii

Learning Deep and Wide: A Spectral Method for Learning Deep Networks

Ling Shao, Di Wu, Xuelong Li
2014 IEEE Transactions on Neural Networks and Learning Systems  
analysis of the best ad-hoc combinatorial joints features distinguish different actions.  ...  Bridging the gap between hand-crafted features and feature learning.  ...  Result: GDBN -a gaussian bernoulli visible layer Deep Belief Network to generate the emission probabilities for hidden markov model. 3DCNN -a 3D Deep Convolutional Neural Networks to generate the emission  ... 
doi:10.1109/tnnls.2014.2308519 pmid:25420251 fatcat:4mnl6tv2xnf3jpzwhp76cvl4ti

Deep Learning for Free-Hand Sketch: A Survey [article]

Peng Xu, Timothy M. Hospedales, Qiyue Yin, Yi-Zhe Song, Tao Xiang, Liang Wang
2022 arXiv   pre-print
The progress of deep learning has immensely benefited free-hand sketch research and applications.  ...  This paper presents a comprehensive survey of the deep learning techniques oriented at free-hand sketch data, and the applications that they enable.  ...  Moreover, matching based on hand-crafted features is inaccurate. Gradually, sketch based 3D model retrieval has been studied within the end-to-end deep learning paradigm [2] , [294] - [299] .  ... 
arXiv:2001.02600v3 fatcat:lek5sivzsrat3i52lqh2eifnia

Deep Reinforcement Learning [article]

Yuxi Li
2018 arXiv   pre-print
Next we discuss RL core elements, including value function, policy, reward, model, exploration vs. exploitation, and representation.  ...  We discuss deep reinforcement learning in an overview style. We draw a big picture, filled with details.  ...  The authors propose policy-space response oracle (PSRO), and its approximation, deep cognitive hierarchies (DCH), to compute best responses to a mixture of policies using deep RL, and to compute new meta-strategy  ... 
arXiv:1810.06339v1 fatcat:kp7atz5pdbeqta352e6b3nmuhy

Deep filter banks for texture recognition, description, and segmentation [article]

Mircea Cimpoi, Subhransu Maji, Iasonas Kokkinos, Andrea Vedaldi
2015 arXiv   pre-print
properties if the convolutional layers of a deep model are used as filter banks.  ...  We obtain in this manner state-of-the-art performance in numerous datasets well beyond textures, an efficient method to apply deep features to image regions, as well as benefit in transferring features  ...  models; (iii) combining FC and FV pooling has a modest benefit and there is no benefit in integrating SIFT features; (iv) in very deep models, most of the performance gain is realized in the very last  ... 
arXiv:1507.02620v2 fatcat:hy7bumxlbvgtdlariuukwu5bqy

Towards Structured Prediction in Bioinformatics with Deep Learning [article]

Yu Li
2020 arXiv   pre-print
Firstly, we can combine deep learning with other classic algorithms, such as probabilistic graphical models, which model the problem structure explicitly.  ...  models can lead to unsatisfactory results.  ...  Instead, previous studies rely heavily on manually crafted features, and consider feature extraction and classification as two separate problems.  ... 
arXiv:2008.11546v1 fatcat:5in2a642b5cj3lweuynl7sniaa

Cross-Lingual Information Retrieval and Semantic Interoperability for Cultural Heritage Repositories

Johanna Monti, Mario Monteleone, Maria Pia di Buono, Federica Marano
2013 Recent Advances in Natural Language Processing  
This paper describes a computational linguistics-based approach for providing interoperability between multi-lingual systems in order to overcome crucial issues like cross-language and cross-collection  ...  " property, (iii) E26 indicates "Physical Feature" class.  ...  Conclusions The proposed architecture ensures not only the coverage of a large knowledge portion but preserves deep semantic relations among different languages.  ... 
dblp:conf/ranlp/MontiMBM13 fatcat:drdggw3qpbfubhv434yoz2i364

A New Approach for Advertising CTR Prediction Based on Deep Neural Network via Attention Mechanism

Qianqian Wang, Fang'ai Liu, Shuning Xing, Xiaohui Zhao
2018 Computational and Mathematical Methods in Medicine  
Our method exploits dimension reduction based on decomposition, takes advantage of the attention mechanism in neural network modelling, and improves FM to make feature interactions contribute differently  ...  Click-through rate prediction is critical in Internet advertising and affects web publisher's profits and advertiser's payment.  ...  Many successful solutions in both industry and academia largely rely on manually crafting combinatorial features [11] , i.e., constructing new features by combining multiple predictor variables, also  ... 
doi:10.1155/2018/8056541 fatcat:hi25slmrnnggxfkb6cqbsovtpi

Semantic Relations and Deep Learning [article]

Vivi Nastase, Stan Szpakowicz
2021 arXiv   pre-print
A new Chapter 5 of the book, by Vivi Nastase and Stan Szpakowicz, discusses relation classification/extraction in the deep-learning paradigm which arose after the first edition appeared.  ...  This form of relation extraction does not scale well beyond the document level because of the combinatorial explosion of entity-mention combinations at such a high level. Jia et al.'  ...  Gormley et al. [2015] compute substructure embeddings h w i = f w i ⊗ v w i , where f w i is a vector of hand-crafted features, and ⊗ is the outer product.  ... 
arXiv:2009.05426v4 fatcat:rmzoalfwcza4nex7pd4u6w7kbe

Deep Reinforcement Learning, a textbook [article]

Aske Plaat
2022 arXiv   pre-print
We describe the foundations, the algorithms and the applications of deep reinforcement learning. We cover the established model-free and model-based methods that form the basis of the field.  ...  They have learned to fly model helicopters and perform aerobatic manoeuvers such as loops and rolls.  ...  Agents learned to run, jump, crouch and turn as the environment required, without explicit reward shaping or other hand-crafted features. For this experiment a distributed version of PPO was used.  ... 
arXiv:2201.02135v2 fatcat:3icsopexerfzxa3eblpu5oal64

Deep Filter Banks for Texture Recognition, Description, and Segmentation

Mircea Cimpoi, Subhransu Maji, Iasonas Kokkinos, Andrea Vedaldi
2016 International Journal of Computer Vision  
properties if the convolutional layers of a deep model are used as Communicated by  ...  Third, we revisit classic texture represenations, including bag-of-visual-words and the Fisher vectors, in the context of deep learning and show that these have excellent efficiency and generalization  ...  features; (iv) in very deep models, most of the performance gain is realized in the very last few layers.  ... 
doi:10.1007/s11263-015-0872-3 pmid:27471340 pmcid:PMC4946812 fatcat:z7sz65gi5nevbgsh2tt3kzdnzi

Deep Learning and Earth Observation to Support the Sustainable Development Goals [article]

Claudio Persello, Jan Dirk Wegner, Ronny Hänsch, Devis Tuia, Pedram Ghamisi, Mila Koeva, Gustau Camps-Valls
2021 arXiv   pre-print
The synergistic combination of deep learning models and Earth observation promises significant advances to support the sustainable development goals (SDGs).  ...  This paper reviews current deep learning approaches for Earth observation data, along with their application towards monitoring and achieving the SDGs most impacted by the rapid development of deep learning  ...  The results show that an ensemble non-linear regression model, combining the results of the CNN and models based on hand-crafted and GIS features, can explain 75% of the variation in the poverty index  ... 
arXiv:2112.11367v1 fatcat:7eve5dr45vcublfqyzzrccuvxa
« Previous Showing results 1 — 15 out of 718 results