Topic: HADRON - processing AI computations from untapped power of phones&PCs (Read 120 times)
February 23, 2018, 04:06:09 AM
i started mining with my echelon lvl 3 and try to increase my level. hope this post will help.
February 20, 2018, 06:33:05 PM
Hadron lets people use their personal devices—smartphones, laptops, tablets—to perform distributed AI computations in their web browsers. We call this mining. In exchange for performing this work, Hadron (and its customers) reward these miners with Hadron coin.
One of the biggest benefits to this approach is that they can democratize mining in a way that other projects are unable to. Miners don't need to invest in expensive, high-powered mining rigs when mining on the Hadron network.
Even Chromebooks, RPi, NvidiaShield & Smartmirrors(!) are capable of mining with profit. Although having a GPU or a more powerful computer can mean more earnings than someone with a slower device.
The mining itself happens in the browser where it is running AI computations for HADRON.cloud customers. That means your phone and computer's web browser will be running useful neural network computations instead of just a repetitive cryptographic hash puzzle. These computations are used for commerce and for worthy charitable causes. Potentially, some of these computations can benefit society, including detecting diseases like cancer in radiology imagery.
In the whitepaper they do mention a few use cases like:
- Automated tagging and organization of photos from smart phones and cameras
- Filtering of user content for inappropriate content
- Automatic social media tagging
- AI-assisted image annotations for accelerating at-home Mechanical Turk-style work
Some other potential use cases might be:
- Distributed computing for scientific research (a la Folding@Home, mapping neural networks in brain cross-sections, annotating patterns in radio telescopy data to identify exoplanets, etc)
- Identifying features in de-identified medical data (CAT scans, etc)
- Performing analysis of traffic data for Google
- Plagiarism checking for academic institutions/publishers
- Signal analysis or processing audio/video.
One of the current applications is classifying images, from which it not a too far leap to detecting copyrighted audio (possibly video) material on streaming sites, or nsfw content for that matter. Platforms such as Youtube still use A LOT of manual labor to enforce their content rules. The current systems in place are easily fooled by flipping the video or slightly modulating the audio.
Hadron is at the moment in the Alpha stage, mining opened 15th of January2018. As the team filters out bugs & manages to scale the application up, they will gradually add new users to the mining platform.
If you're already mining with GPU & perhaps CPU, chances are your RAM is still untouched - Hadron can make use of that.
More info:
https://hadron.cloud/
https://hadron.cloud/whitepaper.html
One of the biggest benefits to this approach is that they can democratize mining in a way that other projects are unable to. Miners don't need to invest in expensive, high-powered mining rigs when mining on the Hadron network.
Even Chromebooks, RPi, NvidiaShield & Smartmirrors(!) are capable of mining with profit. Although having a GPU or a more powerful computer can mean more earnings than someone with a slower device.
The mining itself happens in the browser where it is running AI computations for HADRON.cloud customers. That means your phone and computer's web browser will be running useful neural network computations instead of just a repetitive cryptographic hash puzzle. These computations are used for commerce and for worthy charitable causes. Potentially, some of these computations can benefit society, including detecting diseases like cancer in radiology imagery.
In the whitepaper they do mention a few use cases like:
- Automated tagging and organization of photos from smart phones and cameras
- Filtering of user content for inappropriate content
- Automatic social media tagging
- AI-assisted image annotations for accelerating at-home Mechanical Turk-style work
Some other potential use cases might be:
- Distributed computing for scientific research (a la Folding@Home, mapping neural networks in brain cross-sections, annotating patterns in radio telescopy data to identify exoplanets, etc)
- Identifying features in de-identified medical data (CAT scans, etc)
- Performing analysis of traffic data for Google
- Plagiarism checking for academic institutions/publishers
- Signal analysis or processing audio/video.
One of the current applications is classifying images, from which it not a too far leap to detecting copyrighted audio (possibly video) material on streaming sites, or nsfw content for that matter. Platforms such as Youtube still use A LOT of manual labor to enforce their content rules. The current systems in place are easily fooled by flipping the video or slightly modulating the audio.
Hadron is at the moment in the Alpha stage, mining opened 15th of January2018. As the team filters out bugs & manages to scale the application up, they will gradually add new users to the mining platform.
If you're already mining with GPU & perhaps CPU, chances are your RAM is still untouched - Hadron can make use of that.
More info:
https://hadron.cloud/
https://hadron.cloud/whitepaper.html
Jump to: