Top 10 Arxiv Papers Today
2.523 Mikeys
#1. Gradient Descent Maximizes the Margin of Homogeneous Neural Networks
Kaifeng Lyu, Jian Li
Recent works on implicit regularization have shown that gradient descent
converges to the max-margin direction for logistic regression with one-layer or
multi-layer linear networks. In this paper, we generalize this result to
homogeneous neural networks, including fully-connected and convolutional neural
networks with ReLU or LeakyReLU activations. In particular, we study the
gradient flow (gradient descent with infinitesimal step size) optimizing the
logistic loss or cross-entropy loss of any homogeneous model (possibly
non-smooth), and show that if the training loss decreases below a certain
threshold, then we can define a smoothed version of the normalized margin which
increases over time. We also formulate a natural constrained optimization
problem related to margin maximization, and prove that both the normalized
margin and its smoothed version converge to the objective value at a KKT point
of the optimization problem. Furthermore, we extend the above results to a
large family of loss functions. We conduct several experiments...
more |
pdf
| html
Figures
None.
Tweets
BrundageBot:
Gradient Descent Maximizes the Margin of Homogeneous Neural Networks. Kaifeng Lyu and Jian Li https://t.co/iFhdDqZKgH
arxivml:
"Gradient Descent Maximizes the Margin of Homogeneous Neural Networks",
Kaifeng Lyu, Jian Li
https://t.co/rQ3HtB4iIT
StatsPapers:
Gradient Descent Maximizes the Margin of Homogeneous Neural Networks. https://t.co/Wq8DGus1Zq
Github
Robust evasion attacks against neural network to find adversarial examples
Stargazers: 343
Subscribers: 9
Subscribers: 9
Forks: 102
Open Issues: 3
Open Issues: 3
Youtube
None.
Other stats
Sample Sizes : None.
Authors: 2
Total Words: 19227
Unqiue Words: 3404
2.192 Mikeys
#2. Grounding Object Detections With Transcriptions
Yasufumi Moriya, Ramon Sanabria, Florian Metze, Gareth J. F. Jones
A vast amount of audio-visual data is available on the Internet thanks to
video streaming services, to which users upload their content. However, there
are difficulties in exploiting available data for supervised statistical models
due to the lack of labels. Unfortunately, generating labels for such amount of
data through human annotation can be expensive, time-consuming and prone to
annotation errors. In this paper, we propose a method to automatically extract
entity-video frame pairs from a collection of instruction videos by using
speech transcriptions and videos. We conduct experiments on image recognition
and visual grounding tasks on the automatically constructed entity-video frame
dataset of How2. The models will be evaluated on new manually annotated portion
of How2 dev5 and val set and on the Flickr30k dataset. This work constitutes a
first step towards meta-algorithms capable of automatically construct
task-specific training sets.
more |
pdf
| html
Figures
Tweets
Github
This repository contains code and metadata of How2 dataset
Stargazers: 29
Subscribers: 6
Subscribers: 6
Forks: 3
Open Issues: 0
Open Issues: 0
Youtube
None.
Other stats
Sample Sizes : None.
Authors: 4
Total Words: 5891
Unqiue Words: 1905
2.192 Mikeys
#3. Machine Learning Approach to Earthquake Rupture Dynamics
Sabber Ahamed, Eric G. Daub
Simulating dynamic rupture propagation is challenging due to the
uncertainties involved in the underlying physics of fault slip, stress
conditions, and frictional properties of the fault. A trial and error approach
is often used to determine the unknown parameters describing rupture, but
running many simulations usually requires human review to determine how to
adjust parameter values and is thus not very efficient. To reduce the
computational cost and improve our ability to determine reasonable stress and
friction parameters, we take advantage of the machine learning approach. We
develop two models for earthquake rupture propagation using the artificial
neural network (ANN) and the random forest (RF) algorithms to predict if a
rupture can break a geometric heterogeneity on a fault. We train the models
using a database of 1600 dynamic rupture simulations computed numerically.
Fault geometry, stress conditions, and friction parameters vary in each
simulation. We cross-validate and test the predictive power of the models using
an...
more |
pdf
| html
Figures
None.
Tweets
arxivml:
"Machine Learning Approach to Earthquake Rupture Dynamics",
Sabber Ahamed, Eric G． Daub
https://t.co/MKdmAJOH25
Memoirs:
Machine Learning Approach to Earthquake Rupture Dynamics. https://t.co/OG0LS93W0l
Github
This repository contains python codes to classify earthquake rupture based on random forest and neural network.
Stargazers: 0
Subscribers: 1
Subscribers: 1
Forks: 4
Open Issues: 0
Open Issues: 0
Youtube
None.
Other stats
Sample Sizes : None.
Authors: 2
Total Words: 9626
Unqiue Words: 2640
2.191 Mikeys
#4. A Partially Reversible U-Net for Memory-Efficient Volumetric Image Segmentation
Robin Brügger, Christian F. Baumgartner, Ender Konukoglu
One of the key drawbacks of 3D convolutional neural networks for segmentation
is their memory footprint, which necessitates compromises in the network
architecture in order to fit into a given memory budget. Motivated by the
RevNet for image classification, we propose a partially reversible U-Net
architecture that reduces memory consumption substantially. The reversible
architecture allows us to exactly recover each layer's outputs from the
subsequent layer's ones, eliminating the need to store activations for
backpropagation. This alleviates the biggest memory bottleneck and enables very
deep (theoretically infinitely deep) 3D architectures. On the BraTS challenge
dataset, we demonstrate substantial memory savings. We further show that the
freed memory can be used for processing the whole field-of-view (FOV) instead
of patches. Increasing network depth led to higher segmentation accuracy while
growing the memory footprint only by a very small fraction, thanks to the
partially reversible architecture.
more |
pdf
| html
Figures
Tweets
c_f_baumgartner:
More exciting work from @bmic_eth accepted to @miccai2019! Master student Robin Bruegger used reversible units to develop a class of extremely memory-efficient 3D segmentation networks of almost unlimited depth.
Pre-print: https://t.co/YrOAYVF8UP
Code: https://t.co/6ThhPGHnNK https://t.co/Nz9gQSDfiP
c_f_baumgartner:
The non-PDF link is here https://t.co/n9cZe5h6qO for those who prefer this.
arxiv_cscv:
A Partially Reversible U-Net for Memory-Efficient Volumetric Image Segmentation https://t.co/OZkxjfL2Se
Github
Framework for creating (partially) reversible neural networks with PyTorch
Stargazers: 0
Subscribers: 0
Subscribers: 0
Forks: 0
Open Issues: 0
Open Issues: 0
Youtube
None.
Other stats
Sample Sizes : None.
Authors: 3
Total Words: 3227
Unqiue Words: 1230
2.191 Mikeys
#5. Model Agnostic Dual Quality Assessment for Adversarial Machine Learning and an Analysis of Current Neural Networks and Defenses
Danilo Vasconcellos Vargas, Shashank Kotyan
In adversarial machine learning, there are a huge number of attacks of
various types which makes the evaluation of robustness for new models and
defenses a daunting task. To make matters worse, there is an inherent bias in
attacks and defenses. Here, we organize the problems faced (model dependence,
insufficient evaluation, unreliable adversarial samples and perturbation
dependent results) and propose a dual quality assessment method together with
the concept of robustness levels to tackle them. We validate the dual quality
assessment on state-of-the-art models (WideResNet, ResNet, AllConv, DenseNet,
NIN, LeNet and CapsNet) as well as the current hardest defenses proposed at
ICLR 2018 as well as the widely known adversarial training, showing that
current models and defenses are vulnerable in all levels of robustness.
Moreover, we show that robustness to $L_0$ and $L_\infty$ attacks differ
greatly and therefore duality should be taken into account for a correct
assessment. Interestingly, a by-product of the assessment proposed is a...
more |
pdf
| html
Figures
None.
Tweets
BrundageBot:
Model Agnostic Dual Quality Assessment for Adversarial Machine Learning and an Analysis of Current Neural Networks and Defenses. Danilo Vasconcellos Vargas and Shashank Kotyan https://t.co/QyKqJQPapE
Memoirs:
Model Agnostic Dual Quality Assessment for Adversarial Machine Learning and an Analysis of Current Neural Networks and Defenses. https://t.co/E1OfzzS2Xl
onNewHorizons:
It is hard to evaluate the robustness of DNNs. Too many attacking methods and different results based on different amount of perturbation. We proposed a model agnostic dual assessment and the concept of robustness levels to enable this to happen. Today on https://t.co/2u0XptIbnF https://t.co/CA7nU8k7Ez
onNewHorizons:
It's hard to evaluate the robustness of DNNs. Each attacking method will output a different amount of perturbation and accuracy while many won't work for hybrid models. Proposed Solution: a model agnostic dual assessment and robustness levels. Today on https://t.co/r2FZcxVotw #ML https://t.co/C0UW35lJON
arxivml:
"Model Agnostic Dual Quality Assessment for Adversarial Machine Learning and an Analysis of Current Neural Networks…
https://t.co/zYpQf34h44
Github
None.
Youtube
None.
Other stats
Sample Sizes : None.
Authors: 2
Total Words: 0
Unqiue Words: 0
2.174 Mikeys
#6. Deep Learning Development Environment in Virtual Reality
Kevin C. VanHorn, Meyer Zinn, Murat Can Cobanoglu
Virtual reality (VR) offers immersive visualization and intuitive
interaction. We leverage VR to enable any biomedical professional to deploy a
deep learning (DL) model for image classification. While DL models can be
powerful tools for data analysis, they are also challenging to understand and
develop. To make deep learning more accessible and intuitive, we have built a
virtual reality-based DL development environment. Within our environment, the
user can move tangible objects to construct a neural network only using their
hands. Our software automatically translates these configurations into a
trainable model and then reports its resulting accuracy on a test dataset in
real-time. Furthermore, we have enriched the virtual objects with
visualizations of the model's components such that users can achieve insight
about the DL models that they are developing. With this approach, we bridge the
gap between professionals in different fields of expertise while offering a
novel perspective for model analysis and data interaction. We...
more |
pdf
| html
Figures
None.
Tweets
BrundageBot:
Deep Learning Development Environment in Virtual Reality. Kevin C. VanHorn, Meyer Zinn, and Murat Can Cobanoglu https://t.co/COIYv3pIwf
arxivml:
"Deep Learning Development Environment in Virtual Reality",
Kevin C． VanHorn, Meyer Zinn, Murat Can Cobanoglu
https://t.co/0SB2uEn4Ea
Memoirs:
Deep Learning Development Environment in Virtual Reality. https://t.co/MTWZgVPcf3
colourspeak:
RT @yapp1e: Deep Learning Development Environment in Virtual Reality. (arXiv:1906.05925v1 [cs.LG]) https://t.co/fBqSFhQ9Rg
Virtual reality…
FScarfato:
RT @yapp1e: Deep Learning Development Environment in Virtual Reality. (arXiv:1906.05925v1 [cs.LG]) https://t.co/fBqSFhQ9Rg
Virtual reality…
Github
None.
Youtube
None.
Other stats
Sample Sizes : None.
Authors: 3
Total Words: 0
Unqiue Words: 0
2.168 Mikeys
#7. Towards Compact and Robust Deep Neural Networks
Vikash Sehwag, Shiqi Wang, Prateek Mittal, Suman Jana
Deep neural networks have achieved impressive performance in many
applications but their large number of parameters lead to significant
computational and storage overheads. Several recent works attempt to mitigate
these overheads by designing compact networks using pruning of connections.
However, we observe that most of the existing strategies to design compact
networks fail to preserve network robustness against adversarial examples. In
this work, we rigorously study the extension of network pruning strategies to
preserve both benign accuracy and robustness of a network. Starting with a
formal definition of the pruning procedure, including pre-training, weights
pruning, and fine-tuning, we propose a new pruning method that can create
compact networks while preserving both benign accuracy and robustness. Our
method is based on two main insights: (1) we ensure that the training
objectives of the pre-training and fine-tuning steps match the training
objective of the desired robust model (e.g., adversarial...
more |
pdf
| html
Figures
None.
Tweets
BrundageBot:
Towards Compact and Robust Deep Neural Networks. Vikash Sehwag, Shiqi Wang, Prateek Mittal, and Suman Jana https://t.co/MOoqvfqQ5P
arxivml:
"Towards Compact and Robust Deep Neural Networks",
Vikash Sehwag, Shiqi Wang, Prateek Mittal, Suman Jana
https://t.co/BeJPnhuqBn
StatsPapers:
Towards Compact and Robust Deep Neural Networks. https://t.co/IhcCJLX5J2
arxiv_cscv:
Towards Compact and Robust Deep Neural Networks https://t.co/uniKUL70Vy
Github
None.
Youtube
None.
Other stats
Sample Sizes : None.
Authors: 4
Total Words: 7662
Unqiue Words: 2085
2.167 Mikeys
#8. Galaxy-Galaxy Lensing in HSC: Validation Tests and the Impact of Heterogeneous Spectroscopic Training Sets
Joshua S. Speagle, Alexie Leauthaud, Song Huang, Christopher P. Bradshaw, Felipe Ardila, Peter L. Capak, Daniel J. Eisenstein, Daniel C. Masters, Rachel Mandelbaum, Surhud More, Melanie Simet, Cristóbal Sifón
Although photometric redshifts (photo-z's) are crucial ingredients for
current and upcoming large-scale surveys, the high-quality spectroscopic
redshifts currently available to train, validate, and test them are
substantially non-representative in both magnitude and color. We investigate
the nature and structure of this bias by tracking how objects from a
heterogeneous training sample contribute to photo-z predictions as a function
of magnitude and color, and illustrate that the underlying redshift
distribution at fixed color can evolve strongly as a function of magnitude. We
then test the robustness of the galaxy-galaxy lensing signal in 120 deg$^2$ of
HSC-SSP DR1 data to spectroscopic completeness and photo-z biases, and find
that their impacts are sub-dominant to current statistical uncertainties. Our
methodology provides a framework to investigate how spectroscopic
incompleteness can impact photo-z-based weak lensing predictions in future
surveys such as LSST and WFIRST.
more |
pdf
| html
Figures
Tweets
Github
A photometric redshift monstrosity
Stargazers: 9
Subscribers: 8
Subscribers: 8
Forks: 2
Open Issues: 0
Open Issues: 0
Youtube
None.
Other stats
Sample Sizes : None.
Authors: 12
Total Words: 14516
Unqiue Words: 3508
2.152 Mikeys
#9. NLProlog: Reasoning with Weak Unification for Question Answering in Natural Language
Leon Weber, Pasquale Minervini, Jannes Münchmeyer, Ulf Leser, Tim Rocktäschel
Rule-based models are attractive for various tasks because they inherently
lead to interpretable and explainable decisions and can easily incorporate
prior knowledge. However, such systems are difficult to apply to problems
involving natural language, due to its linguistic variability. In contrast,
neural models can cope very well with ambiguity by learning distributed
representations of words and their composition from data, but lead to models
that are difficult to interpret. In this paper, we describe a model combining
neural networks with logic programming in a novel manner for solving multi-hop
reasoning tasks over natural language. Specifically, we propose to use a Prolog
prover which we extend to utilize a similarity function over pretrained
sentence encoders. We fine-tune the representations for the similarity function
via backpropagation. This leads to a system that can apply rule-based reasoning
to natural language, and induce domain-specific rules from training data. We
evaluate the proposed system on two different...
more |
pdf
| html
Figures
Tweets
BrundageBot:
NLProlog: Reasoning with Weak Unification for Question Answering in Natural Language. Leon Weber, Pasquale Minervini, Jannes Münchmeyer, Ulf Leser, and Tim Rocktäschel https://t.co/IIktK7e24C
Github
Neural Logic Reasoning for Question Answering
Stargazers: 1
Subscribers: 1
Subscribers: 1
Forks: 0
Open Issues: 0
Open Issues: 0
Youtube
None.
Other stats
Sample Sizes : None.
Authors: 5
Total Words: 7223
Unqiue Words: 2473
2.148 Mikeys
#10. Fast Radio Bursts: An Extragalactic Enigma
James M. Cordes, Shami Chatterjee
We summarize our understanding of millisecond radio bursts from an
extragalactic population of sources. FRBs occur at an extraordinary rate,
thousands per day over the entire sky with radiation energy densities at the
source about ten billion times larger than those from Galactic pulsars. We
survey FRB phenomenology, source models and host galaxies, coherent radiation
models, and the role of plasma propagation effects in burst detection. The FRB
field is guaranteed to be exciting: new telescopes will expand the sample from
the current $\sim 80$ unique burst sources (and a few secure localizations and
redshifts) to thousands, with burst localizations that enable host-galaxy
redshifts emerging directly from interferometric surveys.
* FRBs are now established as an extragalactic phenomenon.
* Only a few sources are known to repeat. Despite the failure to redetect
other FRBs, they are not inconsistent with all being repeaters.
* FRB sources may be new, exotic kinds of objects or known types in extreme
circumstances. Many...
more |
pdf
| html
Figures
None.
Tweets
higgsinocat:
Fast Radio Bursts: An Extragalactic Enigma. (arXiv:1906.05878v1 [astro-ph.HE]) relevance:72% https://t.co/F9aPx1Pnqu #darkmatter https://t.co/ufcgUQkd8q
StarshipBuilder:
Fast Radio Bursts: An Extragalactic Enigma
https://t.co/tHpcFt6f6s
scimichael:
Fast Radio Bursts: An Extragalactic Enigma
https://t.co/E1x85qW0gG
anomalistnews:
RT @StarshipBuilder: Fast Radio Bursts: An Extragalactic Enigma
https://t.co/tHpcFt6f6s
KlbTheScientist:
RT @StarshipBuilder: Fast Radio Bursts: An Extragalactic Enigma
https://t.co/tHpcFt6f6s
Github
None.
Youtube
None.
Other stats
Sample Sizes : None.
Authors: 2
Total Words: 0
Unqiue Words: 0
About
Assert is a website where the best academic papers on arXiv (computer science, math, physics), bioRxiv (biology), BITSS (reproducibility), EarthArXiv (earth science), engrXiv (engineering), LawArXiv (law), PsyArXiv (psychology), SocArXiv (social science), and SportRxiv (sport research) bubble to the top each day.
Papers are scored (in real-time) based on how verifiable they are (as determined by their Github repos) and how interesting they are (based on Twitter).
To see top papers, follow us on twitter @assertpub_ (arXiv), @assert_pub (bioRxiv), and @assertpub_dev (everything else).
To see beautiful figures extracted from papers, follow us on Instagram.
Tracking 143,632 papers.
Search
Sort results based on if they are interesting or reproducible.
Interesting
Reproducible