This AI can predict how long your relationship will last

This AI can predict how long your relationship will last

  • But researchers from the University of Southern California have trained one to give troubled relationships a death date — which might be the key to saving them.
  • A machine learning algorithm listened in on the therapy sessions of 134 couples, and the researchers fed it information on the lifespan of each relationship.
  • According to the paper: – – While it might sound bleak to think an AI can hear you talk to your sweetheart and be able to determine you’ll break up in three months, there’s another way of looking at its efforts — namely, that hearing its predictions might give couples…
  • I don’t claim to be an expert on relationships, but negative tone of voice can imply a lack of respect or regard — an implication that can be poison for a relationship.
  • So if you’re in therapy and an AI attuned to tone of voice gives your couplehood a death sentence, perhaps that can be a concrete way of correcting an otherwise ephemeral issue.

Researchers from the USC have trained an AI to give troubled relationships a death date — which might be the key to saving them.
Continue reading “This AI can predict how long your relationship will last”

GitHub

Official #TensorFlow implementation of Dense Transformer Networks

  • In this work, we propose Dense Transformer Networks to apply spatial transformation to semantic prediction tasks.
  • The third and fourth rows are the segmentation results of U-Net and DTN, respectively.
  • max_epoch: how many iterations or steps to train

    test_step: how many steps to perform a mini test or validation

    save_step: how many steps to save the model

    summary_step: how many steps to save the summary

    sampledir: where to store predicted samples, please add a / at the end for convinience

    model_name: the name prefix of saved models

    test_epoch: which step to test or predict

    network_depth: how deep of the U-Net including the bottom layer

    class_num: how many classes.

  • We have conv2d for standard convolutional layer, and ipixel_cl for input pixel convolutional layer proposed in our paper.
  • We have deconv for standard deconvolutional layer, ipixel_dcl for input pixel deconvolutional layer, and pixel_dcl for pixel deconvolutional layer proposed in our paper.

Contribute to dtn development by creating an account on GitHub.
Continue reading “GitHub”

Google Sheets now uses machine learning to help you visualize your data

Google Sheets now uses machine learning to help you visualize your data  #CompBindTech

  • After adding the machine learning-powered “Explore” feature last year, which lets you ask natural language questions about your data, it’s now expanding this feature to also automatically build charts for you.
  • All of this is backed by the same natural language understanding tech that already powered the “Explore” feature.
  • It’s worth noting that the previous version of “Explore” could already build graphs for you, but those focused on your complete data set.
  • With this new version, Google also is making it easier to keep in sync data from Sheets that you use in Docs or Slides.
  • You could already update charts you copy into Docs and Slides with just a click, but now you also can do the same with tables.

Google Sheets is getting smarter today. After adding the machine learning-powered “Explore” feature last year, which lets you ask natural language questions..
Continue reading “Google Sheets now uses machine learning to help you visualize your data”

Actress Kristen Stewart (yes, that Kristen Stewart) just released a research paper on artificial intelligence — Quartz

  • Kristen Stewart (yes, that Kristen Stewart) just released a research paper on artificial intelligence
  • Trying to direct the algorithm into producing an artistically satisfying image proved more difficult than expected, according to the paper.
  • (You can see the images in the paper [pdf]).
  • Hollywood has used the concept of artificial intelligence in films for decades.
  • The paper describes the filmmaker’s experiments with style transfer, a popular use of machine learning that transforms one image into the artistic technique and color profile of another.

Hollywood has used the concept of artificial intelligence in films for decades. Now, one A-lister is trying to use AI to make art, instead of just inspire it. Kristen Stewart is most well-known for her star role in the massively successful (and massively mocked) Twilight movies. Less well-known is her interest in AI, laid out in a new paper on…
Continue reading “Actress Kristen Stewart (yes, that Kristen Stewart) just released a research paper on artificial intelligence — Quartz”

Lab41 Reading Group: Swapout: Learning an Ensemble of Deep Architectures

Ensemble of Deep Architectures. #BigData #DeepLearning #MachineLearning #DataScience

  • Swapout samples from every possible stochastic depth and ResNet architecture, both including and not include dropout!
  • When training the network is constantly mutating as units pick different ways of behaving, but at inference time that network needs to be roughly static so that the same input will always yield the same prediction.
  • Stochastic depth can be thought of as randomly selecting between the outcomes {X, F(X)} for each block, so that every unit in the block returns X or F(X) together.
  • Each box represents a block, and each circle is a unit from the block.
  • A unit that is dropped half the time would (by chance) appear in half randomly generated versions of the network.

Next up for the reading group is a paper about a new stochastic training method: Swapout.
Continue reading “Lab41 Reading Group: Swapout: Learning an Ensemble of Deep Architectures”

Doing ad hoc big data analytics? Don’t throw the data away when you’re finished, warns IBM

Doing ad hoc #BigData #Analytics?  #DataScience #MachineLearning #IoT #AI

  • Don’t throw the data away when you’re finished, warns IBM
  • Organisations and lines-of-business have been warned not to casually throw away data after performing ad hoc big data analytics as, increasingly, they may need to refer to the decisions made as a result…
  • Access your subscription from outside of the office
  • To find out how to overcome cloud and infrastructure obstacles, download this review now.
  • The Computing Cloud & Infrastructure Review, explores infrastructure, data centres, and cloud market growth.

IBM’s Alex Chen warns that regulatory requirements means that big data projects will increasingly need to be closely managed
Continue reading “Doing ad hoc big data analytics? Don’t throw the data away when you’re finished, warns IBM”

Google’s AI reasons its way around the London Underground : Nature News & Comment

Google-owned DeepMind learns to navigate London Underground thanks to neural networking  #AI

  • A neural network learns by strengthening connections between virtual neuron-like units.
  • The system performed better than ‘recurrent neural networks’, which also have a memory, but one that is in the fabric of the network itself, and so is less flexible than an external memory.
  • The combination allows the neural network not only to learn, but to use memory to store and recall facts to make inferences like a conventional algorithm.
  • In a paper published in Nature on 12 October 1 , the Google-owned company DeepMind in London reveals that it has taken a step towards overcoming this hurdle by creating a neural network with an external memory.
  • Although the puzzles tackled by DeepMind’s AI are simple, Bengio sees the paper as a signal that neural networks are advancing beyond mere pattern recognition to human-like tasks such as reasoning. “

DeepMind’s latest technique uses external memory to solve tasks that require logic and reasoning — a step toward more human-like AI.
Continue reading “Google’s AI reasons its way around the London Underground : Nature News & Comment”

The brightest minds in artificial intelligence think the government should start taking notes — Quartz

The brightest minds in #DeepLearning #AI think the government should start taking notes

  • Artificial intelligence agents can handle money, or drive our cars, or give legal advice-all this has been done within a framework conceived when computers couldn’t decide for themselves.
  • Any jobs that AI might create are beyond the authors’ imagination.
  • In 2014, Stanford University launched the One Hundred Year Study, a long-term look into the future of artificial intelligence set to publish a paper every five years.
  • “There is no clear definition of AI,” the study reads.
  • But the paper strongly suggests that since artificial intelligence is so widespread and manifests in so many forms, any widespread ruling or central government office to regulate it would be ill-advised.

In 2014, Stanford University launched the One Hundred Year Study, a long-term look into the future of artificial intelligence set to publish a paper every five years. Just two years in, the team released its first report Sept. 1, Artificial Intelligence and Life in 2030. The document outlines the history of AI and where its…
Continue reading “The brightest minds in artificial intelligence think the government should start taking notes — Quartz”

Artificial Intelligence System Predicts Human Interactions – News Center

Hug or handshake? @MIT researchers created an #AI system that can predict human interactions

  • “I’m excited to see how much better the algorithms get if we can feed them a lifetime’s worth of videos,” says Vondrick. “
  • When predicting which of the four actions the person would perform one second later, the algorithm correctly predicted the action more than 43 percent of the time – and humans who have been watching TV for years were only able to predict the next action with 71 percent accuracy.
  • In their second study, the algorithm was shown frames from a video and asked it to predict what object will appear five seconds later.
  • Using a Tesla K40 GPU with the cuDNN -accelerated Caffe deep learning framework, the researchers trained their network on 600 hours of prime-time television shows including The Office and Desperate Housewives .
  • Researchers from MIT’s Computer Science and Artificial Intelligence Laboratory developed an algorithm that can predict whether two individuals will hug, kiss, shake hands or slap five in the next scene.

Read the full article, click here.


@GPUComputing: “Hug or handshake? @MIT researchers created an #AI system that can predict human interactions”


Predicting what will happen in the future is challenging. Researchers from MIT’s Computer Science and Artificial Intelligence Laboratory developed an algorithm that can predict whether two individuals will hug, kiss, shake hands or slap five in the next scene.


Artificial Intelligence System Predicts Human Interactions – News Center