recentpopularlog in

pierredv : machine-learning   14

Radio Frequency-Activity Modeling and Pattern Recognition (RF-AMPR) | 2018
"OBJECTIVE: The PMW 120 Program Office desires a Radio Frequency Activity Modeling and Pattern Recognition (RF-AMPR) capability to perform pattern recognition, anomaly detection, and improved clustering of radio frequency (RF) signals. Specifically, it shall consist of a parametric RF classifier, a generative model of activity in the local electromagnetic environment, a machine learning-based anomaly detection method, and an RF data-clustering algorithm that classifies data that would otherwise be discarded by the parametric classifier."

"DESCRIPTION: Current automated RF data analysis and information discovery methods necessitate discarding significant volumes of sensor data as “non-analyzable”. This SBIR topic seeks to apply machine learning methodologies to better characterize this discarded data, enabling a more complete understanding of RF activity present in a specific environment."

"Anomaly classification shall include “known unknowns”, radio frequency events that are outliers of known classes, and “unknown unknowns”, anomalous RF events that represent new devices or activities."
SBIR  DoD  RF  spectrum  machine-learning  anomaly-classification  ML  AI 
october 2018 by pierredv
AI researchers allege that machine learning is alchemy | Science | AAAS, May 2018
Via John Helm

"Speaking at an AI conference, Rahimi charged that machine learning algorithms, in which computers learn through trial and error, have become a form of "alchemy." Researchers, he said, do not know why some algorithms work and others don't, nor do they have rigorous criteria for choosing one AI architecture over another. Now, in a paper presented on 30 April at the International Conference on Learning Representations in Vancouver, Canada, Rahimi and his collaborators document examples of what they see as the alchemy problem and offer prescriptions for bolstering AI's rigor."

"The issue is distinct from AI's reproducibility problem... It also differs from the "black box" or "interpretability" problem in machine learning"

"Without deep understanding of the basic tools needed to build and train new algorithms, he says, researchers creating AIs resort to hearsay, like medieval alchemists."
ML  machine-learning  AI  ScienceMag 
may 2018 by pierredv
Deep learning - Yann LeCun, Yoshua Bengio & Geoffrey Hinton | Nature May 2015
Deep learning allows computational models that are composed of multiple processing layers to learn representations of data with multiple levels of abstraction. These methods have dramatically improved the state-of-the-art in speech recognition, visual object recognition, object detection and many other domains such as drug discovery and genomics. Deep learning discovers intricate structure in large data sets by using the backpropagation algorithm to indicate how a machine should change its internal parameters that are used to compute the representation in each layer from the representation in the previous layer. Deep convolutional nets have brought about breakthroughs in processing images, video, speech and audio, whereas recurrent nets have shone light on sequential data such as text and speech.
machine-learning  ML  AI  NatureJournal 
april 2018 by pierredv
Machine Learning for Performance Prediction in Mobile Cellular Networks - IEEE Computational Intelligence Magazine ( Volume: 13, Issue: 1, Feb. 2018 )
Janne Riihijarvi ; Petri Mahonen

In this paper, we discuss the application of machine learning techniques for performance prediction problems in wireless networks. These problems often involve using existing measurement data to predict network performance where direct measurements are not available. We explore the performance of existing machine learning algorithms for these problems and propose a simple taxonomy of main problem categories. As an example, we use an extensive real-world drive test data set to show that classical machine learning methods such as Gaussian process regression, exponential smoothing of time series, and random forests can yield excellent prediction results. Applying these methods to the management of wireless mobile networks has the potential to significantly reduce operational costs while simultaneously improving user experience. We also discuss key challenges for future work, especially with the focus on practical deployment of machine learning techniques for performance prediction in mobile wireless networks.
machine-learning  ML  automation  AI  IEEE  cellular  spectrum 
april 2018 by pierredv
A neural network that keeps seeing art where we see mundane objects | Aeon Videos
When mundane objects such as cords, keys and cloths are fed into a live webcam, a machine-learning algorithm ‘sees’ brilliant colours and images such as seascapes and flowers instead. The London-based, Turkish-born visual artist Memo Akten applies algorithms to the webcam feed as a way to reflect on the technology and, by extension, on ourselves. Each instalment in his Learning to See series features a pre-trained deep-neural network ‘trying to make sense of what it sees, in context of what it’s seen before’. In Gloomy Sunday, the algorithm draws from tens of thousands of images scraped from the Google Arts Project, an extensive collection of super-high-resolution images of notable artworks. Set to the voice of the avant-garde singer Diamanda Galás, the resulting video has unexpected pathos, prompting reflection on how our minds construct images based on prior inputs, and not on precise recreations of the outside world.
AeonMagazine  machine-learning  images  image-recognition 
april 2018 by pierredv
CRFS publishes White Paper on Machine Learning — CRFS, Dec 2017
Via Dale Hatfield, March 2018

"Machine Learning and RF Spectrum Intelligence Gathering"

"Many applications that are central to RF spectrum intelligence gathering require some sort of pattern recognition. For example, to classify a signal by type we need to identify the particular pattern associated with the modulation, while to recognise that there is an interesting signal present in received data, we need to distinguish between pattern and noise.

In this White Paper, we explore how machine learning techniques can be applied to these applications of signal classification and anomaly detection to deliver faster and more effective performance to customers."
CRFS  signal-processing  enforcement  AI  ML  machine-learning  spectrum 
april 2018 by pierredv
The Radio Frequency Spectrum + Machine Learning = A New Wave in Radio Technology
"The radio frequency spectrum is becoming increasingly crowded and a new DARPA program will examine how leading-edge machine learning can help understand all the signals in the crowd"

“What I am imagining is the ability of an RF Machine Learning system to see and understand the composition of the radio frequency spectrum – the kinds of signals occupying it, differentiating those that are ‘important’ from the background, and identifying those that don’t follow the rules,” said Tilghman.

The RFMLS program features four technical components that would integrate into future RFML systems:

Feature Learning: ...
Attention and Saliency: ...
Autonomous RF Sensor Configuration: ...
Waveform Synthesis: ...
DARPA  AI  RF  spectrum  machine-learning 
november 2017 by pierredv
Why should I trust you? Explaining the predictions of any classifier | the morning paper
Summary of “Why Should I Trust You? Explaining the Predictions of Any Classifier Ribeiro et al., KDD 2016
machine-learning  AI  automation 
november 2017 by pierredv
I-Connect007 :: Article The Radio Frequency Spectrum + Machine Learning = a New Wave in Radio Technology
" there now is a need to apply ML to the invisible realm of radio frequency (RF) signals, according to program manager Paul Tilghman of DARPA’s Microsystems Technology Office. To further that cause, DARPA today announced its new Radio Frequency Machine Learning Systems (RFMLS) program."

“What I am imagining is the ability of an RF Machine Learning system to see and understand the composition of the radio frequency spectrum – the kinds of signals occupying it, differentiating those that are ‘important’ from the background, and identifying those that don’t follow the rules,” said Tilghman.

"stand up an RF forensics capability to identify unique and peculiar signals amongst the proverbial cocktail party of signals out there"

"The RFMLS program features four technical components that would integrate into future RFML systems:"
= feature learning
= attention and saliency
= autonomous RF sensor configuration
= waveform synthesis
machine-learning  ML  DARPA  enforcement  forensics  mirror-worlds  *  spectrum 
september 2017 by pierredv
Smart buildings predict when critical systems are about to fail | New Scientist issue 3110, Jan 2017
"They trained a machine learning algorithm on data from the first half of 2015, looking for differences in the readings of similar appliances. They then tested it on data from the second half of the year – could it predict faults before they happened? The system predicted 76 out of 124 real faults, including 41 out of 44 where an appliance's temperature rose above tolerable levels, with a false positive rate of 5 per cent"


"Finnish start-up Leanheat puts a wireless temperature, humidity and pressure sensor into apartments to remotely control heating and monitor appliance health. Its system is now installed in nearly 400 apartment blocks, says chief executive Jukka Aho."Once we had these sensors in place, very quickly there was evidence that buildings were not controlled optimally," he says. Instead of adjusting heating simply based on the outside temperature, Leanheat's models take into account how the weather has changed. Has the temperature fallen to zero from 10 degrees – or risen from 10 below?"
"US-based start-up Augury is installing acoustic sensors in machines to listen for audible changes in function and spot potentially imminent failures. CEO Saar Yoskovitz says Augury has 'diagnosed' machines in hospitals, power plants, data centres and a university campus."
NewScientist  building  architecture  RF-MirrorWorlds  prediction  AI  machine-learning 
may 2017 by pierredv
DARPA | Spectrum Collaboration Challenge
The DARPA Spectrum Collaboration Challenge (SC2) is the first-of-its-kind collaborative machine-learning competition to overcome scarcity in the radio frequency (RF) spectrum. Today, spectrum is managed by dividing it into rigid, exclusively licensed bands. This human-driven process is not adaptive to the dynamics of supply and demand, and thus cannot exploit the full potential capacity of the spectrum. In SC2, competitors will reimagine a new, more efficient wireless paradigm in which radio networks autonomously collaborate to dynamically determine how the spectrum should be used moment to moment.
DARPA  spectrum  challenge  machine-learning 
october 2016 by pierredv
Statistical and causal approaches to machine learning - YouTube
Published on Dec 16, 2014 "Where would you take machine learning? 2014's Milner Award winner Professor Bernhard Schölkopf, Max Planck Institute for Intelligent Systems, talks through basic concepts of machine learning to pioneering research now widely used in science and industry."
RoyalSociety  statistics  causality  video  machine-learning  lectures 
december 2014 by pierredv

Copy this bookmark:

to read