Filters








11,544 Hits in 4.3 sec

Extractor-Based Time-Space Lower Bounds for Learning [article]

Sumegha Garg, Ran Raz, Avishay Tal
2017 arXiv   pre-print
Our proof builds on [R17] that gave a general technique for proving memory-samples lower bounds.  ...  a tight Ω(( |X|) · ( |A|)) lower bound on the size of the memory, rather than a bound of Ω({( |X|)^2,( |A|)^2}) obtained in previous works [R17,MM17b].  ...  Acknowledgement We would like to thank Pooya Hatami and Avi Wigderson for very helpful conversations.  ... 
arXiv:1708.02639v1 fatcat:uaecyps4cvcgjddesycf5dtava

Computational Complexity of Discrete Problems (Dagstuhl Seminar 19121)

Anna Gál, Rahul Santhanam, Till Tantau, Michael Wagner
2019 Dagstuhl Reports  
For the natural setting of d = Θ(log n), our result implies an Ω(lg 2 n) lower bound, which is a quadratic improvement over the highest (non-oblivious) cell-probe lower bound for ANN.  ...  This is the first super-logarithmic unconditional lower bound for ANN against general (non black-box) data structures.  ...  Executive Summary Anna Gál, Overview of Talks Planarity, Exclusivity, and Unambiguity We develop an extension of recent analytic methods for obtaining time-space tradeoff lower bounds for problems of learning  ... 
doi:10.4230/dagrep.9.3.64 dblp:journals/dagstuhl-reports/GalST19 fatcat:gypvueu2kreclllv3n6eaz6ch4

Neural Fuzzy Extractors: A Secure Way to Use Artificial Neural Networks for Biometric User Authentication [article]

Abhishek Jana, Md Kamruzzaman Sarker, Monireh Ebrahimi, Pascal Hitzler, George T Amariucai
2020 arXiv   pre-print
The NFE thus offers all the performance advantages of modern deep-learning-based classifiers, and all the security of standard fuzzy extractors.  ...  In this paper, we introduce a secure way to handle user-specific information involved with the use of vector-space classifiers or artificial neural networks for biometric authentication.  ...  It would be possible in this framework to decide (at least approximately, based on the upper and lower bounds on the entropy of a user's biometric data) whether fingerprintbased biometrics, for instance  ... 
arXiv:2003.08433v1 fatcat:ntqrft3ig5bztiitwthja5odye

Adversarial Sampling for Active Learning

Christoph Mayer, Radu Timofte
2020 2020 IEEE Winter Conference on Applications of Computer Vision (WACV)  
This paper proposes ASAL, a new GAN based active learning method that generates high entropy samples.  ...  Instead of directly annotating the synthetic samples, ASAL searches similar samples from the pool and includes them for training.  ...  (ii) Random sampling (lower bound, baseline) and (iii) the fully supervised model (upper bound). In addition we report for a subset of the experiments the results of Core-set based AL (MNIST & SVHN).  ... 
doi:10.1109/wacv45572.2020.9093556 dblp:conf/wacv/0007T20 fatcat:dsqtf76kt5cfjehsnb2pzvgchy

Reinforcement Learning Based Optimal Camera Placement for Depth Observation of Indoor Scenes [article]

Yichuan Chen and Manabu Tsukada and Hiroshi Esaki
2021 arXiv   pre-print
learning backbone and a feature extractor to extract features from the observed point cloud layer-by-layer.  ...  To this problem, an OCP online solution to depth observation of indoor scenes based on reinforcement learning is proposed in this paper.  ...  CONCLUSION In this study, a reinforcement learning-based OCP system for depth observation tasks in indoor scenes was proposed.  ... 
arXiv:2110.11106v1 fatcat:y4oqeoc77fcezavrqclt5f3xfm

Deep Private-Feature Extraction

Seyed Ali Osia, Ali Taheri, Ali Shahin Shamsabadi, Minos Katevas, Hamed Haddadi, Hamid R. R. Rabiee
2018 IEEE Transactions on Knowledge and Data Engineering  
We present and evaluate Deep Private-Feature Extractor (DPFE), a deep model which is trained and evaluated based on information theoretic constraints.  ...  Our results on benchmark image datasets demonstrate that under moderate resource utilization, DPFE can achieve high accuracy for primary tasks while preserving the privacy of sensitive features.  ...  as the lower bound for mutual information.  ... 
doi:10.1109/tkde.2018.2878698 fatcat:ksmlcrwkkrbexg7c4x3yfy4pmu

Memory-Sample Lower Bounds for Learning Parity with Noise [article]

Sumegha Garg, Pravesh K. Kothari, Pengda Liu, Ran Raz
2021 arXiv   pre-print
In fact, we study memory-sample lower bounds for a large class of learning problems, as characterized by [GRT'18], when the samples are noisy.  ...  In this work, we show, for the well-studied problem of learning parity under noise, where a learner tries to learn x=(x_1,...  ...  Acknowledgements We would like to thank Avishay Tal and Greg Valiant for the helpful discussions.  ... 
arXiv:2107.02320v1 fatcat:dond4gfr5vfodiqhvsowqp66pa

Joint 3D Proposal Generation and Object Detection from View Aggregation [article]

Jason Ku, Melissa Mozifian, Jungwook Lee, Ali Harakeh, Steven Waslander
2018 arXiv   pre-print
in 3D space.  ...  for deployment on autonomous vehicles.  ...  Feature Extractor: We compare the detection results of our feature extractor to that of the base VGG-based feature extractor proposed by MV3D.  ... 
arXiv:1712.02294v4 fatcat:hdz4n3dkcncjlm63yui4asadku

Is a Good Representation Sufficient for Sample Efficient Reinforcement Learning? [article]

Simon S. Du and Sham M. Kakade and Ruosong Wang and Lin F. Yang
2020 arXiv   pre-print
These lower bounds highlight that having a good (value-based, model-based, or policy-based) representation in and of itself is insufficient for efficient reinforcement learning, unless the quality of this  ...  Furthermore, our lower bounds also imply exponential separations on the sample complexity between 1) value-based learning with perfect representation and value-based learning with a good-but-not-perfect  ...  ACKNOWLEDGMENTS The authors would like to thank Yuping Luo, Wenlong Mou, Martin Wainwright, Mengdi Wang and Yifan Wu for insightful discussions.  ... 
arXiv:1910.03016v4 fatcat:k4swyw2myfaflmvwwtwhofnxmu

Optimized feature space learning for generating efficient binary codes for image retrieval

Abin Jose, Erik Stefan Ottlik, Christian Rohlfing, Jens-Rainer Ohm
2022 Signal processing : image communication 100  
In this paper, a novel approach for learning a low-dimensional optimized feature space for image retrieval with minimum intra-class variance and maximum inter-class variance is proposed.  ...  Our experiments prove that we could train the neural network to reach the theoretical lower bound of loss corresponding to the negative sum of the correlation coefficients.  ...  Conclusions and future work In this paper, we propose a novel deep learning based approach for learning an optimized feature space for image retrieval.  ... 
doi:10.18154/rwth-2022-01115 fatcat:j4ommasfvfhszkjrmx4bgllxhq

Deep Private-Feature Extraction [article]

Seyed Ali Osia, Ali Taheri, Ali Shahin Shamsabadi, Kleomenis Katevas, Hamed Haddadi, Hamid R. Rabiee
2018 arXiv   pre-print
We present and evaluate Deep Private-Feature Extractor (DPFE), a deep model which is trained and evaluated based on information theoretic constraints.  ...  Our results on benchmark image datasets demonstrate that under moderate resource utilization, DPFE can achieve high accuracy for primary tasks while preserving the privacy of sensitive features.  ...  Lower bound for I(f; z). We derive a variational lower bound for mutual information by first expressing Lemma 1 and then proving Theorem 2. Lemma 1.  ... 
arXiv:1802.03151v2 fatcat:lwpa4buybbbm5l6rem7p6gguaq

Automatic Fish Age Determination across Different Otolith Image Labs Using Domain Adaptation

Alba Ordoñez, Line Eikvil, Arnt-Børre Salberg, Alf Harbitz, Bjarki Þór Elvarsson
2022 Fishes  
the way for requiring less human effort and availability of expertise by means of deep learning (DL).  ...  We conclude that careful consideration must be given before DL-based predictors are applied to perform large scale inference.  ...  Acknowledgments: We thank the Researach Council of Norway for funding, but also Auður Súsanna Bjarnadóttir and Sigurlína Gunnarsdóttir from the Marine and Freshwater Research Institute for their involvement  ... 
doi:10.3390/fishes7020071 fatcat:4ajfjdb5anf7nmgaxkwywq7naa

SPARK: Static Program Analysis Reasoning and Retrieving Knowledge [article]

Wasuwee Sodsong, Bernhard Scholz, Sanjay Chawla
2017 arXiv   pre-print
In this work, we present a machine learning pipeline that induces a security analyzer for programs by example.  ...  The security analyzer determines whether a program is either secure or insecure based on symbolic rules that were deduced by our machine learning pipeline.  ...  However, the computation of the metric requires double exponential time. We propose a sampling-based similarity comparison that computes the lower bound in polynomial time complexity.  ... 
arXiv:1711.01024v1 fatcat:edmkrvffjjhbtpkim46xkrrsza

Fast Tracking-by-Detection of Bus Passengers with Siamese CNNs

Claire Labit-Bonis, Jerpme Thomas, Frederic Lerasle, Francisco Madrigal
2019 2019 16th IEEE International Conference on Advanced Video and Signal Based Surveillance (AVSS)  
Evaluations on our own large scale in-situ dataset are very promising in terms of performances and real-time constraint expected for on-board processing.  ...  trajectories over time.  ...  Because it offers good performance and an online, real-time approach for multi-object tracking (MOT) based on CNN embedding for similarity learning, the work of Wojke et al. is in complete agreement with  ... 
doi:10.1109/avss.2019.8909843 dblp:conf/avss/Labit-BonisTLM19 fatcat:n46bui6v7bafdl5be7apo474au

Simple and effective localized attribute representations for zero-shot learning [article]

Shiqi Yang, Kai Wang, Luis Herranz, Joost van de Weijer
2021 arXiv   pre-print
Our method can be implemented easily, which can be used as a new baseline for zero shot-learning. In addition, our localized representations are highly interpretable as attribute-specific heatmaps.  ...  Some recent papers have shown the importance of localized features together with fine-tuning the feature extractor to obtain discriminative and transferable features.  ...  Acknowledgements We acknowledge the support from Huawei Kirin Solution, the European Union's H2020 research under the Marie Sklodowska-Curie grant agreement No.665919 and the Spanish Government funding for  ... 
arXiv:2006.05938v3 fatcat:d7e3tqwv4rbypnhjgmg5z5frqq
« Previous Showing results 1 — 15 out of 11,544 results