id
string
sources
list
title
string
abstract
string
authors
list
categories
list
fields_of_study
list
published_date
timestamp[s]
url
string
pdf_url
string
arxiv_id
string
doi
string
citation_count
int64
influential_citation_count
int64
has_code
bool
code_url
string
venue
string
quality_score
float64
eadf6c9c394be4bddd136cbdd22aeb4bfa060d637186ae3491a97dc68738b88b
[ "arxiv", "semantic_scholar" ]
Catastrophic forgetting: still a problem for DNNs
We investigate the performance of DNNs when trained on class-incremental visual problems consisting of initial training, followed by retraining with added visual classes. Catastrophic forgetting (CF) behavior is measured using a new evaluation procedure that aims at an application-oriented view of incremental learning....
[ "B. Pfülb", "A. Gepperth", "S. Abdullah", "A. Kilian" ]
[ "cs.LG", "cs.CV" ]
[ "Computer Science" ]
2019-05-20T00:00:00
https://arxiv.org/abs/1905.08077
https://arxiv.org/pdf/1905.08077v1
1905.08077
10.1007/978-3-030-01418-6_48
25
1
false
null
International Conference on Artificial Neural Networks
0.3537
e8c445b9d0c94a1430d378f7aac1993bc592d155ef21abe15d11589df56794aa
[ "arxiv", "semantic_scholar" ]
Alpha MAML: Adaptive Model-Agnostic Meta-Learning
Model-agnostic meta-learning (MAML) is a meta-learning technique to train a model on a multitude of learning tasks in a way that primes the model for few-shot learning of new tasks. The MAML algorithm performs well on few-shot learning problems in classification, regression, and fine-tuning of policy gradients in reinf...
[ "Harkirat Singh Behl", "Atılım Güneş Baydin", "Philip H. S. Torr" ]
[ "cs.LG", "cs.AI", "stat.ML" ]
[ "Computer Science", "Mathematics" ]
2019-05-17T00:00:00
https://arxiv.org/abs/1905.07435
https://arxiv.org/pdf/1905.07435v1
1905.07435
null
72
0
false
null
arXiv.org
0.4658
082cba3250274d6083c0bb86e58265294263f8f0b4c823ff38c1fb10cc27471a
[ "arxiv", "semantic_scholar" ]
TapNet: Neural Network Augmented with Task-Adaptive Projection for Few-Shot Learning
Handling previously unseen tasks after given only a few training examples continues to be a tough challenge in machine learning. We propose TapNets, neural networks augmented with task-adaptive projection for improved few-shot learning. Here, employing a meta-learning strategy with episode-based training, a network and...
[ "Sung Whan Yoon", "Jun Seo", "Jaekyun Moon" ]
[ "cs.LG", "stat.ML" ]
[ "Computer Science", "Mathematics" ]
2019-05-16T00:00:00
https://arxiv.org/abs/1905.06549
https://arxiv.org/pdf/1905.06549v2
1905.06549
null
301
18
false
null
International Conference on Machine Learning
0.6394
5c02d84e68caf25f598335610edc7aa2fcc39b062657f50763cd58e53df026e0
[ "arxiv", "semantic_scholar" ]
A Neural Network-Evolutionary Computational Framework for Remaining Useful Life Estimation of Mechanical Systems
This paper presents a framework for estimating the remaining useful life (RUL) of mechanical systems. The framework consists of a multi-layer perceptron and an evolutionary algorithm for optimizing the data-related parameters. The framework makes use of a strided time window to estimate the RUL for mechanical component...
[ "David Laredo", "Zhaoyin Chen", "Oliver Schütze", "Jian-Qiao Sun" ]
[ "cs.LG", "cs.NE", "stat.ML" ]
[ "Computer Science", "Mathematics", "Medicine" ]
2019-05-15T00:00:00
https://arxiv.org/abs/1905.05918
https://arxiv.org/pdf/1905.05918v1
1905.05918
10.1016/j.neunet.2019.04.016
54
0
false
null
Neural Networks
0.4351
14812c009f5d54c4f935f395d9d7ee0328a0b521dc122e25f4616d78aa56e296
[ "arxiv", "semantic_scholar" ]
Embeddings and Representation Learning for Structured Data
Performing machine learning on structured data is complicated by the fact that such data does not have vectorial form. Therefore, multiple approaches have emerged to construct vectorial representations of structured data, from kernel and distance approaches to recurrent, recursive, and convolutional neural networks. Re...
[ "Benjamin Paaßen", "Claudio Gallicchio", "Alessio Micheli", "Alessandro Sperduti" ]
[ "cs.LG", "stat.ML" ]
[ "Computer Science", "Mathematics" ]
2019-05-15T00:00:00
https://arxiv.org/abs/1905.06147
https://arxiv.org/pdf/1905.06147v1
1905.06147
null
8
0
false
null
The European Symposium on Artificial Neural Networks
0.2386
579d2c3beafcec2368f9388660d630eb629837e1ec23db6fc887bb61d6a892ef
[ "arxiv", "semantic_scholar" ]
Learning Generative Models across Incomparable Spaces
Generative Adversarial Networks have shown remarkable success in learning a distribution that faithfully recovers a reference distribution in its entirety. However, in some cases, we may want to only learn some aspects (e.g., cluster or manifold structure), while modifying others (e.g., style, orientation or dimension)...
[ "Charlotte Bunne", "David Alvarez-Melis", "Andreas Krause", "Stefanie Jegelka" ]
[ "cs.LG", "stat.ML" ]
[ "Computer Science", "Mathematics" ]
2019-05-14T00:00:00
https://arxiv.org/abs/1905.05461
https://arxiv.org/pdf/1905.05461v2
1905.05461
10.3929/ETHZ-B-000382654
119
7
false
null
International Conference on Machine Learning
0.5198
5ff1a020dabc8c4b4303a2940c0804a72ae739dfed1099bd9fde0cef68b75c12
[ "arxiv", "semantic_scholar" ]
Fast and Reliable Architecture Selection for Convolutional Neural Networks
The performance of a Convolutional Neural Network (CNN) depends on its hyperparameters, like the number of layers, kernel sizes, or the learning rate for example. Especially in smaller networks and applications with limited computational resources, optimisation is key. We present a fast and efficient approach for CNN a...
[ "Lukas Hahn", "Lutz Roese-Koerner", "Klaus Friedrichs", "Anton Kummert" ]
[ "cs.CV", "cs.LG" ]
[ "Computer Science" ]
2019-05-06T00:00:00
https://arxiv.org/abs/1905.01924
https://arxiv.org/pdf/1905.01924v1
1905.01924
null
0
0
false
null
The European Symposium on Artificial Neural Networks
0
595392074956175759cd3d497d5f1e86e0c41b4d1c169e9040d8d40b2179ff2d
[ "arxiv", "semantic_scholar" ]
Unsupervised Representation Learning with Minimax Distance Measures
We investigate the use of Minimax distances to extract in a nonparametric way the features that capture the unknown underlying patterns and structures in the data. We develop a general-purpose and computationally efficient framework to employ Minimax distances with many machine learning methods that perform on numerica...
[ "Morteza Haghir Chehreghani" ]
[ "cs.LG", "cs.AI", "stat.ML" ]
[ "Computer Science", "Mathematics" ]
2019-04-27T00:00:00
https://arxiv.org/abs/1904.13223
https://arxiv.org/pdf/1904.13223v3
1904.13223
10.1007/s10994-020-05886-4
14
1
false
null
Machine-mediated learning
0.294
3e9bd46f4be18aa89a38f0dd89329b85639be35f39d44d4b8e5f44951a77c110
[ "arxiv", "semantic_scholar" ]
Facilitating Bayesian Continual Learning by Natural Gradients and Stein Gradients
Continual learning aims to enable machine learning models to learn a general solution space for past and future tasks in a sequential manner. Conventional models tend to forget the knowledge of previous tasks while learning a new task, a phenomenon known as catastrophic forgetting. When using Bayesian models in continu...
[ "Yu Chen", "Tom Diethe", "Neil Lawrence" ]
[ "cs.LG", "cs.AI", "stat.ML" ]
[ "Computer Science", "Mathematics" ]
2019-04-24T00:00:00
https://arxiv.org/abs/1904.10644
https://arxiv.org/pdf/1904.10644v1
1904.10644
null
15
2
false
null
arXiv.org
0.301
1cb9a9ce979b3482fee857ad5402ced1877e1ce99d5667454038ff3c0e4bbddb
[ "arxiv", "semantic_scholar" ]
Continual Learning with Self-Organizing Maps
Despite remarkable successes achieved by modern neural networks in a wide range of applications, these networks perform best in domain-specific stationary environments where they are trained only once on large-scale controlled data repositories. When exposed to non-stationary learning environments, current neural netwo...
[ "Pouya Bashivan", "Martin Schrimpf", "Robert Ajemian", "Irina Rish", "Matthew Riemer", "Yuhai Tu" ]
[ "cs.NE" ]
[ "Computer Science" ]
2019-04-19T00:00:00
https://arxiv.org/abs/1904.09330
https://arxiv.org/pdf/1904.09330v1
1904.09330
null
17
0
false
null
arXiv.org
0.3138
fdc836043ed8ecd0768fd4db515dff29c2f66cf329b020cdefc67e5dafc9cb87
[ "arxiv", "semantic_scholar" ]
Three scenarios for continual learning
Standard artificial neural networks suffer from the well-known issue of catastrophic forgetting, making continual or lifelong learning difficult for machine learning. In recent years, numerous methods have been proposed for continual learning, but due to differences in evaluation protocols it is difficult to directly c...
[ "Gido M. van de Ven", "Andreas S. Tolias" ]
[ "cs.LG", "cs.AI", "cs.CV", "stat.ML" ]
[ "Computer Science", "Mathematics" ]
2019-04-15T00:00:00
https://arxiv.org/abs/1904.07734
https://arxiv.org/pdf/1904.07734v1
1904.07734
null
1,057
79
false
null
arXiv.org
0.9515
48a2e72678690b407ffe680449b22b8333b8511c044200551c9614e7e668679d
[ "arxiv", "semantic_scholar" ]
Transfer Learning with Sparse Associative Memories
In this paper, we introduce a novel layer designed to be used as the output of pre-trained neural networks in the context of classification. Based on Associative Memories, this layer can help design Deep Neural Networks which support incremental learning and that can be (partially) trained in real time on embedded devi...
[ "Quentin Jodelet", "Vincent Gripon", "Masafumi Hagiwara" ]
[ "cs.LG", "cs.CV", "stat.ML" ]
[ "Computer Science", "Mathematics" ]
2019-04-04T00:00:00
https://arxiv.org/abs/1904.02420
https://arxiv.org/pdf/1904.02420v3
1904.02420
10.1007/978-3-030-30487-4_39
0
0
false
null
International Conference on Artificial Neural Networks
0
6985f1e254553076b92eebf33f8ad47453098cb8957d14b7c32ca7a22f22311a
[ "arxiv", "semantic_scholar" ]
Learn to Grow: A Continual Structure Learning Framework for Overcoming Catastrophic Forgetting
Addressing catastrophic forgetting is one of the key challenges in continual learning where machine learning systems are trained with sequential or streaming tasks. Despite recent remarkable progress in state-of-the-art deep learning, deep neural networks (DNNs) are still plagued with the catastrophic forgetting proble...
[ "Xilai Li", "Yingbo Zhou", "Tianfu Wu", "Richard Socher", "Caiming Xiong" ]
[ "cs.LG", "cs.CV" ]
[ "Computer Science", "Mathematics" ]
2019-03-31T00:00:00
https://arxiv.org/abs/1904.00310
https://arxiv.org/pdf/1904.00310v3
1904.00310
null
525
16
false
null
International Conference on Machine Learning
0.6802
cc8e1bed4d2dd901fcf87784e8c17dd87adbb0ce2cc1a4c302ba2e74c4085c19
[ "arxiv", "semantic_scholar" ]
On-line learning dynamics of ReLU neural networks using statistical physics techniques
We introduce exact macroscopic on-line learning dynamics of two-layer neural networks with ReLU units in the form of a system of differential equations, using techniques borrowed from statistical physics. For the first experiments, numerical solutions reveal similar behavior compared to sigmoidal activation researched ...
[ "Michiel Straat", "Michael Biehl" ]
[ "cs.LG", "cond-mat.dis-nn", "stat.ML" ]
[ "Computer Science", "Physics", "Mathematics" ]
2019-03-18T00:00:00
https://arxiv.org/abs/1903.07378
https://arxiv.org/pdf/1903.07378v1
1903.07378
null
11
0
false
null
The European Symposium on Artificial Neural Networks
0.2698
2a28853e157a7139e2a487d0006fd719f3e513b09b2e5d52936e6c4443987e59
[ "arxiv", "semantic_scholar" ]
Communication-Efficient Federated Deep Learning with Asynchronous Model Update and Temporally Weighted Aggregation
Federated learning obtains a central model on the server by aggregating models trained locally on clients. As a result, federated learning does not require clients to upload their data to the server, thereby preserving the data privacy of the clients. One challenge in federated learning is to reduce the client-server c...
[ "Yang Chen", "Xiaoyan Sun", "Yaochu Jin" ]
[ "cs.LG", "cs.AI", "cs.DC", "stat.ML" ]
[ "Computer Science", "Mathematics", "Medicine" ]
2019-03-18T00:00:00
https://arxiv.org/abs/1903.07424
https://arxiv.org/pdf/1903.07424v1
1903.07424
10.1109/TNNLS.2019.2953131
527
28
false
null
IEEE Transactions on Neural Networks and Learning Systems
0.7312
8c42a0cebcdd399fed324369a7fc5d72d7bc69c26140e624c375caff4abca227
[ "arxiv", "semantic_scholar" ]
Continual Learning in Practice
This paper describes a reference architecture for self-maintaining systems that can learn continually, as data arrives. In environments where data evolves, we need architectures that manage Machine Learning (ML) models in production, adapt to shifting data distributions, cope with outliers, retrain when necessary, and ...
[ "Tom Diethe", "Tom Borchert", "Eno Thereska", "Borja Balle", "Neil Lawrence" ]
[ "stat.ML", "cs.LG" ]
[ "Mathematics", "Computer Science" ]
2019-03-12T00:00:00
https://arxiv.org/abs/1903.05202
https://arxiv.org/pdf/1903.05202v2
1903.05202
null
78
5
false
null
Neural Information Processing Systems
0.4744
a9c86f17ad94c38deb380c073f1bb6afc2b7a99ece866a4659cd98b6f23816a4
[ "arxiv", "semantic_scholar" ]
Continual Learning via Neural Pruning
We introduce Continual Learning via Neural Pruning (CLNP), a new method aimed at lifelong learning in fixed capacity models based on neuronal model sparsification. In this method, subsequent tasks are trained using the inactive neurons and filters of the sparsified network and cause zero deterioration to the performanc...
[ "Siavash Golkar", "Michael Kagan", "Kyunghyun Cho" ]
[ "cs.LG", "cs.NE", "q-bio.NC", "stat.ML" ]
[ "Computer Science", "Biology", "Mathematics" ]
2019-03-11T00:00:00
https://arxiv.org/abs/1903.04476
https://arxiv.org/pdf/1903.04476v1
1903.04476
null
181
10
false
null
arXiv.org
0.565
1a9a3535c46cebee31e28ad7fd39f9761baa271e8c45975afac5df4d99f68a7c
[ "arxiv", "semantic_scholar" ]
Complementary Learning for Overcoming Catastrophic Forgetting Using Experience Replay
Despite huge success, deep networks are unable to learn effectively in sequential multitask learning settings as they forget the past learned tasks after learning new tasks. Inspired from complementary learning systems theory, we address this challenge by learning a generative model that couples the current task to the...
[ "Mohammad Rostami", "Soheil Kolouri", "Praveen K. Pilly" ]
[ "cs.LG", "stat.ML" ]
[ "Computer Science", "Mathematics" ]
2019-03-11T00:00:00
https://arxiv.org/abs/1903.04566
https://arxiv.org/pdf/1903.04566v2
1903.04566
10.24963/ijcai.2019/463
82
1
false
null
International Joint Conference on Artificial Intelligence
0.4798
69305eb94930e06b6f7fb0b02c6706b4a82086da4a2ba2061ee43568837d87a2
[ "arxiv", "semantic_scholar" ]
Interpolation Consistency Training for Semi-Supervised Learning
We introduce Interpolation Consistency Training (ICT), a simple and computation efficient algorithm for training Deep Neural Networks in the semi-supervised learning paradigm. ICT encourages the prediction at an interpolation of unlabeled points to be consistent with the interpolation of the predictions at those points...
[ "Vikas Verma", "Kenji Kawaguchi", "Alex Lamb", "Juho Kannala", "Arno Solin", "Yoshua Bengio", "David Lopez-Paz" ]
[ "stat.ML", "cs.AI", "cs.LG" ]
[ "Computer Science", "Medicine", "Mathematics" ]
2019-03-09T00:00:00
https://arxiv.org/abs/1903.03825
https://arxiv.org/pdf/1903.03825v5
1903.03825
10.1016/j.neunet.2021.10.008
911
105
false
null
International Joint Conference on Artificial Intelligence
1
5f3b16e3874ed26bd53f48886b51b43c8b56084760f4dd5d057d031496f24a2e
[ "arxiv", "semantic_scholar" ]
Transfer Learning Using Ensemble Neural Networks for Organic Solar Cell Screening
Organic Solar Cells are a promising technology for solving the clean energy crisis in the world. However, generating candidate chemical compounds for solar cells is a time-consuming process requiring thousands of hours of laboratory analysis. For a solar cell, the most important property is the power conversion efficie...
[ "Arindam Paul", "Dipendra Jha", "Reda Al-Bahrani", "Wei-keng Liao", "Alok Choudhary", "Ankit Agrawal" ]
[ "cs.LG", "physics.chem-ph", "stat.ML" ]
[ "Computer Science", "Physics", "Mathematics" ]
2019-03-07T00:00:00
https://arxiv.org/abs/1903.03178
https://arxiv.org/pdf/1903.03178v4
1903.03178
10.1109/IJCNN.2019.8852446
22
0
false
null
IEEE International Joint Conference on Neural Network
0.3404
4764c15a5844e41937eec236a617e3bd4af382fe4860672b36e4c6cb0598e85a
[ "arxiv", "semantic_scholar" ]
PDP: A General Neural Framework for Learning Constraint Satisfaction Solvers
There have been recent efforts for incorporating Graph Neural Network models for learning full-stack solvers for constraint satisfaction problems (CSP) and particularly Boolean satisfiability (SAT). Despite the unique representational power of these neural embedding models, it is not clear how the search strategy in th...
[ "Saeed Amizadeh", "Sergiy Matusevych", "Markus Weimer" ]
[ "cs.LG", "cs.LO", "cs.NE", "stat.ML" ]
[ "Computer Science", "Mathematics" ]
2019-03-05T00:00:00
https://arxiv.org/abs/1903.01969
https://arxiv.org/pdf/1903.01969v1
1903.01969
null
23
3
false
null
arXiv.org
0.3451
c52d6cc088146c56d993eafc17c4b2d87dfd3123c121493079e8eb99c153a1ac
[ "arxiv", "semantic_scholar" ]
Scalable and Order-robust Continual Learning with Additive Parameter Decomposition
While recent continual learning methods largely alleviate the catastrophic problem on toy-sized datasets, some issues remain to be tackled to apply them to real-world problem domains. First, a continual learning model should effectively handle catastrophic forgetting and be efficient to train even with a large number o...
[ "Jaehong Yoon", "Saehoon Kim", "Eunho Yang", "Sung Ju Hwang" ]
[ "cs.LG", "stat.ML" ]
[ "Computer Science", "Mathematics" ]
2019-02-25T00:00:00
https://arxiv.org/abs/1902.09432
https://arxiv.org/pdf/1902.09432v3
1902.09432
null
209
20
false
null
International Conference on Learning Representations
0.6611
e70ab08b1fa983c335ffd2a37a9f8eb013e74096d323fbcd32088efdd9359ac3
[ "arxiv", "semantic_scholar" ]
Deep Bayesian Multi-Target Learning for Recommender Systems
With the increasing variety of services that e-commerce platforms provide, criteria for evaluating their success become also increasingly multi-targeting. This work introduces a multi-target optimization framework with Bayesian modeling of the target events, called Deep Bayesian Multi-Target Learning (DBMTL). In this f...
[ "Qi Wang", "Zhihui Ji", "Huasheng Liu", "Binqiang Zhao" ]
[ "cs.LG", "cs.IR", "stat.ML" ]
[ "Computer Science", "Mathematics" ]
2019-02-25T00:00:00
https://arxiv.org/abs/1902.09154
https://arxiv.org/pdf/1902.09154v1
1902.09154
null
14
0
false
null
arXiv.org
0.294
87e6f5f778c6c9d5b97c59bcd76177dec6f12328e87f630311628dda440fe4f2
[ "arxiv", "semantic_scholar" ]
Differentially Private Continual Learning
Catastrophic forgetting can be a significant problem for institutions that must delete historic data for privacy reasons. For example, hospitals might not be able to retain patient data permanently. But neural networks trained on recent data alone will tend to forget lessons learned on old data. We present a differenti...
[ "Sebastian Farquhar", "Yarin Gal" ]
[ "stat.ML", "cs.LG" ]
[ "Mathematics", "Computer Science" ]
2019-02-18T00:00:00
https://arxiv.org/abs/1902.06497
https://arxiv.org/pdf/1902.06497v1
1902.06497
null
12
1
false
null
arXiv.org
0.2785
81fab21e758913394f5a79a92a7435887e4f080ef4663d3b624063d37e254589
[ "arxiv", "semantic_scholar" ]
A Unifying Bayesian View of Continual Learning
Some machine learning applications require continual learning - where data comes in a sequence of datasets, each is used for training and then permanently discarded. From a Bayesian perspective, continual learning seems straightforward: Given the model posterior one would simply use this as the prior for the next task....
[ "Sebastian Farquhar", "Yarin Gal" ]
[ "stat.ML", "cs.LG" ]
[ "Mathematics", "Computer Science" ]
2019-02-18T00:00:00
https://arxiv.org/abs/1902.06494
https://arxiv.org/pdf/1902.06494v1
1902.06494
null
80
7
false
null
arXiv.org
0.4771
d904892f020e2e45c11f27e6cb00fb01337d6a95791c834a4be5a2df37ecf5c6
[ "arxiv", "semantic_scholar" ]
Scaling Limits of Wide Neural Networks with Weight Sharing: Gaussian Process Behavior, Gradient Independence, and Neural Tangent Kernel Derivation
Several recent trends in machine learning theory and practice, from the design of state-of-the-art Gaussian Process to the convergence analysis of deep neural nets (DNNs) under stochastic gradient descent (SGD), have found it fruitful to study wide random neural networks. Central to these approaches are certain scaling...
[ "Greg Yang" ]
[ "cs.NE", "cond-mat.dis-nn", "cs.LG", "math-ph", "stat.ML" ]
[ "Computer Science", "Physics", "Mathematics" ]
2019-02-13T00:00:00
https://arxiv.org/abs/1902.04760
https://arxiv.org/pdf/1902.04760v3
1902.04760
null
311
36
false
null
arXiv.org
0.7841
106f6a9ebe3ffdedaa5d5af5565185ad0f20edbda8549eadff0a1cf8d91aa1e8
[ "arxiv", "semantic_scholar" ]
Controlled Forgetting: Targeted Stimulation and Dopaminergic Plasticity Modulation for Unsupervised Lifelong Learning in Spiking Neural Networks
Stochastic gradient descent requires that training samples be drawn from a uniformly random distribution of the data. For a deployed system that must learn online from an uncontrolled and unknown environment, the ordering of input samples often fails to meet this criterion, making lifelong learning a difficult challeng...
[ "Jason M. Allred", "Kaushik Roy" ]
[ "cs.NE", "cs.LG" ]
[ "Computer Science", "Medicine" ]
2019-02-08T00:00:00
https://arxiv.org/abs/1902.03187
https://arxiv.org/pdf/1902.03187v2
1902.03187
10.3389/fnins.2020.00007
38
4
false
null
Frontiers in Neuroscience
0.3978
79b5a8b6b873d83ad123b3eca4bf14a5e1d76a1f4543b819c5e863ff8d3cba62
[ "arxiv", "semantic_scholar" ]
On ADMM in Deep Learning: Convergence and Saturation-Avoidance
In this paper, we develop an alternating direction method of multipliers (ADMM) for deep neural networks training with sigmoid-type activation functions (called \textit{sigmoid-ADMM pair}), mainly motivated by the gradient-free nature of ADMM in avoiding the saturation of sigmoid-type activations and the advantages of ...
[ "Jinshan Zeng", "Shao-Bo Lin", "Yuan Yao", "Ding-Xuan Zhou" ]
[ "cs.LG", "math.OC", "stat.ML" ]
[ "Computer Science", "Mathematics" ]
2019-02-06T00:00:00
https://arxiv.org/abs/1902.02060
https://arxiv.org/pdf/1902.02060v3
1902.02060
null
38
3
false
null
Journal of machine learning research
0.3978
92023f6ecc2141ec54bb624b3d940a31f0dcb2a83ccb4bcc8aeee8589c247f87
[ "arxiv", "semantic_scholar" ]
Generalisation dynamics of online learning in over-parameterised neural networks
Deep neural networks achieve stellar generalisation on a variety of problems, despite often being large enough to easily fit all their training data. Here we study the generalisation dynamics of two-layer neural networks in a teacher-student setup, where one network, the student, is trained using stochastic gradient de...
[ "Sebastian Goldt", "Madhu S. Advani", "Andrew M. Saxe", "Florent Krzakala", "Lenka Zdeborová" ]
[ "stat.ML", "cond-mat.dis-nn", "cond-mat.stat-mech", "cs.LG" ]
[ "Computer Science", "Mathematics", "Physics" ]
2019-01-25T00:00:00
https://arxiv.org/abs/1901.09085
https://arxiv.org/pdf/1901.09085v1
1901.09085
null
15
0
false
null
arXiv.org
0.301
1ad24db00ba0caf229ef7fd0954100ba486e66f7bc4fbea36e105d579cadd032
[ "arxiv", "semantic_scholar" ]
Unsupervised Learning of Neural Networks to Explain Neural Networks (extended abstract)
This paper presents an unsupervised method to learn a neural network, namely an explainer, to interpret a pre-trained convolutional neural network (CNN), i.e., the explainer uses interpretable visual concepts to explain features in middle conv-layers of a CNN. Given feature maps of a conv-layer of the CNN, the explaine...
[ "Quanshi Zhang", "Yu Yang", "Ying Nian Wu" ]
[ "cs.LG", "cs.AI", "stat.ML" ]
[ "Computer Science", "Mathematics" ]
2019-01-21T00:00:00
https://arxiv.org/abs/1901.07538
https://arxiv.org/pdf/1901.07538v1
1901.07538
null
1
0
false
null
arXiv.org
0.0753
8c815760bbedba307ef530e4b189758ccd7995c56ecf61034fe7200add06d0f7
[ "arxiv", "semantic_scholar" ]
Hierarchical Attentional Hybrid Neural Networks for Document Classification
Document classification is a challenging task with important applications. The deep learning approaches to the problem have gained much attention recently. Despite the progress, the proposed models do not incorporate the knowledge of the document structure in the architecture efficiently and not take into account the c...
[ "Jader Abreu", "Luis Fred", "David Macêdo", "Cleber Zanchettin" ]
[ "cs.CL", "cs.AI", "cs.LG", "cs.NE" ]
[ "Computer Science" ]
2019-01-20T00:00:00
https://arxiv.org/abs/1901.06610
https://arxiv.org/pdf/1901.06610v2
1901.06610
10.1007/978-3-030-30493-5_39
38
5
false
null
International Conference on Artificial Neural Networks
0.3978