Fundamental Resource Trade-offs for Encoded Distributed Optimization

arXiv.org Machine Learning

Dealing with the shear size and complexity of today's massive data sets requires computational platforms that can analyze data in a parallelized and distributed fashion. A major bottleneck that arises in such modern distributed computing environments is that some of the worker nodes may run slow. These nodes a.k.a.~stragglers can significantly slow down computation as the slowest node may dictate the overall computational time. A recent computational framework, called encoded optimization, creates redundancy in the data to mitigate the effect of stragglers. In this paper we develop novel mathematical understanding for this framework demonstrating its effectiveness in much broader settings than was previously understood. We also analyze the convergence behavior of iterative encoded optimization algorithms, allowing us to characterize fundamental trade-offs between convergence rate, size of data set, accuracy, computational load (or data redundancy), and straggler toleration in this framework.


Blade Runner--Autoencoded Whitney Museum of American Art

#artificialintelligence

The artist and computer scientist Terence Broad built an autoencoder, a type of artificial neural network, and showed it the classic science-fiction film Blade Runner (1982). He trained the autoencoder to remember every individual frame of the film and to reconstruct each one as a memory, on view here. In the original film, a bounty hunter hunts down androids that are so well engineered that they are indistinguishable from humans. Here, we face a similar challenge, as we trying to identify the original film within the AI's program's perception of it. Terence Broad, Blade Runner--Autoencoded, 2016 Advance tickets are required.


Credit card number and password encoder / decoder

@machinelearnbot

Here's some simple JavaScript code to encode numbers, such as credit card numbers, passwords made up of digits, phone numbers, social security numbers, dates such as 20131014 etc. Enter number to encode / decode in box, on the web page in question Email the encoded number (it should start with e) to your contact Your contact use the same form, enters the encoded number, select Encrypt / Decrypt, and then the original number is immediately retrieved. Your contact use the same form, enters the encoded number, select Encrypt / Decrypt, and then the original number is immediately retrieved. This code is very simple, it is by no means strong encryption. It is indeed less sophisticated than uuencode. But uuencode is for geeks, while our app is easy to use by any mainstream people.


How to Prepare Text Data for Machine Learning with scikit-learn - Machine Learning Mastery

@machinelearnbot

Text data requires special preparation before you can start using it for predictive modeling. The text must be parsed to remove words, called tokenization. Then the words need to be encoded as integers or floating point values for use as input to a machine learning algorithm, called feature extraction (or vectorization). The scikit-learn library offers easy-to-use tools to perform both tokenization and feature extraction of your text data. In this tutorial, you will discover exactly how you can prepare your text data for predictive modeling in Python with scikit-learn.


[Research Article] Self-assembly of genetically encoded DNA-protein hybrid nanoscale shapes

Science

We characterized the DNA-recognizing domains of the TAL effectors with respect to binding affinity and sequence specificity. To construct the staple proteins, we fused two TAL proteins via a custom peptide linker and tested for the ability to connect two separate double-helical DNA domains. For creating larger objects containing multiple staple protein connections, we identified a set of rules regarding the optimal spacing between these connections. On the basis of these rules, we could create megadalton-scale objects that realize a variety of structural motifs, such as custom curvatures, vertices, and corners. Each of those objects was built from a set of 12 double-TAL staple proteins and a template DNA double strand with designed sequence.