Stream

Join this community to post or comment

HotHardware

AI Research  - 
 
"P-p-p poker face, p-p poker face" #CarnegieMellon #AI #poker
These days, machine-learning along with artificial intelligence are becoming increasingly capable, and we're seeing their use applied to things we might not have immediately expected. Case in point: last month, we talked about how AI was being used to craft the most perfect beer. But as wonderous as AI beer may sound, researchers at Carnegie
3
3
Add a comment...
 
...Now that deep learning has proven so effective with speech recognition, image recognition, and translation, natural language is the next frontier. “In the past, people had to build large lexicons, dictionaries, ontologies,” Suleman says. “But with neural nets, we no longer need to do that. A neural net can learn from raw data.”
https://www.wired.com/2017/01/microsoft-thinks-machines-can-learn-converse-chats-become-game/
Microsoft just bought a startup that's embracing an AI technique typically used to master games---so it can teach machines to carry on a conversation.
9
Add a comment...

Boris Kazachenko

AI Research  - 
 
What’s wrong with CNNs: dirty similarity to coarse kernels
CNN is the most successful method now, so I will use it to plug in my approach.

First, similarity measure in CNN is a product between input and kernel.
Why product? Input * kernel obviously overweights close vs. distant input-kernel pairs, for the same value of input + kernel. So, it does detect similarity, but in a dirty way: there is no differentiation between match and difference. Both are included in the product, although match is overrepresented. But difference obviously doesn’t belong in the value of similarity, these are mutually exclusive concepts. I think this accounts for “confirmation bias”, both in CNNs and in a human mind.
I suggest that match (similarity) should be a measure of compression, see part 1 in my intro: www.cognitivealgorithm.info

Second, both input and kernel are arrays, and kernel is formed by many instances of training. Thus, input-to-kernel comparison is far more coarse, thus potentially less selective and efficient than comparison between adjacent / consecutive pixels in my algorithm. Which accounts for abysmal scalability of CNNs.
There are many more theoretically justified distinctions, covered in the link above.

This is an open project, pseudocode: https://docs.google.com/document/d/1GElEvTlwwshZl1Z8KYetMoCPGxfJzq7zKC8G2rll160/edit?usp=sharing
But I am willing to pay for long-term collaboration, please make me an offer if interested.


2
Add a comment...
 
How To Resume Download Failed Files in UC Browser [ FIXED]

Resume Download Failed Files in UC Browser with a simple trick. Thus, now you don't have to think twice before downloading anything from UC browser -

hackerztrickz.com/resume-download-failed-files/
Resume Download Failed Files in UC Browser with a simple trick. Thus, now you don't have to think twice before downloading anything from UC browser.
3
Add a comment...
The FDA approval of a cloud based machine learning application to be used in a clinical setting to help physicians understand how a heart is functioning signals a major breakthrough. Cutting examination time from up to an hour to just 15 seconds paves the way for more AI algorithms in healthcare.
13
2
Add a comment...
 
Google, Facebook, & major news organizations like the New York Times are scrambling to use AI to help fact checkers & identify fake news in print & online. But AI advances in video & audio will soon make it possible to convincingly fake images & voices, too. Read more in my January AI and Humans column:
http://mlmug.org/pdfs/1701.pdf
at pages 49-50
or here:
https://drive.google.com/file/d/0BwR6Uymtc-cmdHh5ZEVGWV81SHc/view?usp=sharing
3
Hendrik Boom's profile photoXavier Williams's profile photo
8 comments
 
Ya google a company that tells us if any government is trying to have info deleted off the internet is going to use this to de-educate us...... Wait that doesn make sense
Add a comment...

Mehrzad Karami

AI Software  - 
 
XNOR.ai is, essentially, a bit of clever computer-native math that enables AI-like models for vision and speech recognition to run practically anywhere. It has the potential to be transformative for the industry....
Code is open to all on Github https://xnor.ai
22
7
illi haike's profile photo
2 comments
 
IPhone comes with one
Add a comment...
 
As a precautionary measure, a group of tech giants have committed $27 million USD to MIT and others for research on the appropriate ethical governance of artificial intelligence.
https://edgylabs.com/2017/01/19/ai-ethics-mit-harvard/ 
Media Lab at MIT and Harvard's Berkman Klein Center have received generous donations from billionaire technology founders to begin research into AI ethics.
5
3
Luke James's profile photo
 
Bad image. I robot was such a horrible adaptation of caves of steel.
Add a comment...
 
Will AI be used in predicting and avoiding disasters? Has any study yet happened?
173 votes  -  votes visible to Public
Yes AI will be used
71%
Not there yet
10%
It is WIP
6%
Would be great to do that
13%
7
Add a comment...
31 comments

About this community

Community formed to discuss ideas and developments in the field of Artificial Intelligence, the discipline of programming computers to do tasks that require intelligence when done by humans. Posting Guidelines: --------------------------- 1. Make sure it has not already been posted. Use the search box on the left to find out. If it has, then comment on the existing post, even if it's old. If you want to post another article on the same topic, include the link as a comment on the existing post. 2. Make sure what you are posting is of high quality and on topic. That means it makes some contribution to discussion, not just rehash common disciplinary knowledge. If you don't have disciplinary knowledge, then lurk for a while do lots of reading and post when you are sure you will make a contribution. Make sure your post does not include grammar or spelling mistakes. 3. If you are posting a link or poll, the link/poll you post must also abide by these guidelines. Add a few lines that makes the argument for why this is being posted here and what the relation to AI is. If you want to drive discussion in a certain direction, ask a question in the context of the link/poll. 4. Posting the same content to many communities is frowned upon. Not only does Google+ flag these as spam automatically, but it shows you are not being sufficiently on topic in your post. 5. This is not the place for marketing your products. If you post about a commercial product it must also include an statement regarding what AI methods are used and how. Any posts that don't fit these guidelines may be removed.

MEGA DROID

General  - 
 
Hello everybody I want to make an A.I custom android ROM and I learned python what else do I need? And do I need to learn neural networks? Thanks in advance.
4
MEGA DROID's profile photo
10 comments
 
+Paul Peacock​ By the way which NLP would you recommend working with from the ones you posted above?
Add a comment...

Mehrzad Karami

General Purpose AI  - 
 
By listening to everything that is happening in your house, as these bots do, they learn how we think, live, work and play. They are gathering massive amounts of data about us. And that raises a dark side of this technology: the privacy risks and possible misuse by technology companies. Neither Amazon nor Google is forthcoming about what it is doing with all of the data it gathers and how it will protect us from hackers who exploit weaknesses in the infrastructure leading to its servers.
https://singularityhub.com/2017/01/07/love-your-bot-but-know-its-always-listening/
In the 2013 movie “Her,” Theodore Twombly, a lonely writer, falls in love with a digital assistant designed to meet his every need. She sorts emails, helps get a book published, provides personal advice and ultimately becomes his girlfriend. The assistant, Samantha, is AI software capable of learning at an astonishing pace. Samantha will remain …
9
Randall Lee Reetz's profile photo
13 comments
 
Ask your AI enhanced world what I'd did for you today? Ask it how it maximized the use of the data it collects on you to make you more productive? Ask it what of your work it go done for you while you slept? NOTHING. SILENCE, CRICKETS.
Add a comment...

Rude Bano

General  - 
13
Randall Lee Reetz's profile photo
 
Exactly how fast are your eyes spinning dude?

Add a comment...

Sharmishtha Sen

AI Research  - 
Artificial intelligence chat bots are easy to write in Python with the AIML package. AIML stands for Artificial Intelligence Markup Language, but it is just simple XML. These code examples will walk you through how to create your own artificial intelligence chat bot using Python.
15
1
Stéphane NUFER's profile photoSharmishtha Sen's profile photo
2 comments
 
+Stéphane NUFER No problem! Glad it helped u! 😊 
Add a comment...
 
Here’s a sentence you don’t get to read everyday: Kristen Stewart has surprised the artificial intelligence community by publishing a paper on machine learning. The Twilight actress recently made...
20
4
Mihai Kulcsar's profile photoSebastian Hoffmann's profile photo
11 comments
 
The paper can be summed up as "We fiddled with the parameters of Neural Style Transfer", and gives no indication of the specific contribution beyond "She provided the painting to sample the style from". The only actual numbers are that they used 256 iterations, that GPU memory was the limiting factor on the process, which is why they used downscaled images for the transfer, then scaled them up and denoised them, and there is also a set of pictures that's maybe helpful to eyeball the style transfer ratio that you're aiming for. And, of course, the references are handy to drill into the literature available on style transfer.
arxiv.org - arxiv.org/pdf/1701.04928v1.pdf
Add a comment...

Karthika Qpt

Robotics  - 
 
Robird is a Flying Robot for controlling Birds at Airports, Harbours and Agriculture Fields
https://www.youtube.com/watch?v=bBc_ODxLJm0
7
Add a comment...
 
 
AI Exonerated
The highway agency found that while Tesla’s Autopilot feature didn’t prevent a crash in Florida, the system performed as it was intended.
30
8
Sebastian Fernandez Alberdi's profile photoFran Marzoa's profile photo
9 comments
 
+Gam3 0ver It's worth to add that the manual of the car is extremely clear about not relying on the autopilot and keep your attention on the road 100% of the time.

Add a comment...
 
+World Economic Forum discussing the AI and its effect on society:

- Data ecosystems for privacy and security
- Algorithms for greater social inclusion
- Institutions to build trust among stakeholders

IMHO the most interesting talk was from Ito's concerning the lack of diversity in the training data set.

https://www.youtube.com/watch?v=rw4aWeL3v5g

17
9
J. Toby Knudsen's profile photoKarl Smithe's profile photo
3 comments
 
Society is going to cope with AI when it only copes with Planned Obsolescence by pretending it hasn't been happening for the last 50 years.  Tell me another one!  How much CO2 is in the atmosphere because of PO.  Now we have throw away computers.  I have gotten a drone and computers out of the garbage and fixed them.  All we have is Simulated Intelligence.
Add a comment...
 
In one experiment, researchers at the Google Brain artificial intelligence research group had software design a machine-learning system to take a test used to benchmark software that processes language. What it came up with surpassed previously published results from software designed by humans...

AI Software Learns to Make AI Software. http://google.com/newsstand/s/CBIw78vkhTQ
Google and others think software that learns to learn could take over some work done by AI experts.
22
12
Sylvain JULIEN's profile photoJeremy Mone (JarethGK)'s profile photo
2 comments
 
Do you want to make a technological singularity people?!
Because this is how you make a technological singularity!

(Keep up the good work! I look forward to working with our robotic overlords soon. Especially in virtual reality...)
Add a comment...