Friday, April 14, 2017

Facebook failed to remove offensive posts, despite its moderators being informed: Reports

Facebook failed to remove offensive posts, despite its moderators being informed: Reports

Image: Reuters
Facebook failed to remove dozens of instances of extremist and child pornography even after the social network’s moderators were directly informed of the potentially illegal content, an investigation by The Times showed on Thursday.
Using a fake profile set up last month, a Times journalist found images and videos glorifying Islamic State and recent deadly attacks in London and Egypt, along with graphic images of child abuse, and asked site moderators to remove them.
Facebook moderators removed some of reported images but left untouched pro-jihadist posts praising recent attacks and calling for new ones. The company appeared to take action only after The Times identified itself as reporting a story on the matter.
Failure to remove content which is illegal under British law after company officials have been notified of its existence could expose Facebook to criminal prosecution for its role in encouraging the publication and distribution of such imagery.
The social media giant faces new laws in countries around the world to force it to move faster to combat illegal content but it has struggled to keep pace as illicit posts can reappear as fast as they are identified and taken down.
A Facebook spokesman said the company had now removed all the images identified by the Times as potentially illegal, acknowledging that they “violate our policies and have no place on Facebook”.
“We are sorry that this occurred,” Facebook Vice President of Operations Justin Osofsky said in a statement. “It is clear that we can do better, and we’ll continue to work hard to live up to the high standards people rightly expect of Facebook.”
A spokesman for London’s Metropolitan Police called for individuals to report extremist content to it via an online form. It declined to comment on whether it was investigating if Facebook failed to act when notified of the illegal content.
“Where material breaches UK terrorism laws, the Counter Terrorism Internet Referral Unit (CTIRU) will, where possible, seek the removal of the content by working with the relevant internet hosting company,” the spokesman said.
Reuters
Publish date: April 14, 2017 11:27 am| Modified date: April 14, 2017 11:27 am

Facebook: About 1.2 billion users actively use Messenger to stay in touch

Facebook: About 1.2 billion users actively use Messenger to stay in touch

Image Credit: Facebook
Over 1.2 billion people use Facebook Messenger every month globally to stay in touch with the people and businesses that matter to them the most. In the last few months, Facebook has added group video calling, a powerful and fast camera – preloaded with stickers, masks, frames, effects and the ability to share photos and videos to your Day, making it easy and fun to share what’s happening in your life, moment by moment.
“I keep on hearing powerful stories about how our product is becoming a more important part of your daily lives. So from all of us here at Messenger, a heartfelt thank you to all of you for giving us a chance to build something good and more meaningful for you,” said David Marcus, Vice President, Messaging Products, Facebook, in a post on Wednesday.
The company has also launched Mentions and Reactions to make group conversations on Messenger fun and engaging. Last month, Facebook Messenger added a new feature Live Location that lets you share – for an hour within a private direct or group message thread – your real-time location on a map.
Publish date: April 14, 2017 11:07 am| Modified date: April 14, 2017 11:07 am

Wednesday, April 12, 2017

GoPro Hero5 Session review: A good action camera for the price but needs improvement

GoPro Hero5 Session review: A good action camera for the price but needs improvement

GoPro Hero5 Session is the latest ‘action’ camera launched by GoPro in the Indian market earlier this year. Session 5 is the mid-tier action camera with the premium GoPro Hero5 Black on the top as the premium offering by the company. The GoPro Hero5 Session is priced at Rs 29,500.
GoPro Hero5 Session
Action cameras are used to shoot sports, bike journeys, among other things that are dangerous for devices like DSLRs or smartphone cameras to shoot. The danger here refers to the equipment getting damaged while biking, swimming or engaging in other types of adventurous sports. This is the first time I have extensively used an action camera as my early encounters were limited to shooting with DSLRs and smartphone cameras.
GoPro Hero5 Session is supposed to be aimed at consumers who want the performance and feature set of GoPro action cameras, but at an affordable price point. So let’s jump in to see how Hero5 Session fares in the review.
Build and Design: 7.5/10
The build quality of Hero5 Session is sturdy and top notch for an action cam. The rubberised top along with solid plastic exterior on the sides works perfectly in securing the internals of the camera.
GoPro Hero5 Session
One thing to note is that you don’t get anything remarkable in terms of design except the form factor. Hero5 Session is shaped like a small cube along with a microSD card slot and USB Type-C slot to charge in through a push door on the left side from the back.
GoPro Hero5 Session
The GoPro Hero5 Session does not come with a microSD card out of the box and you need to use a high-speed card, Class 10 or more for smooth 4K recording.
GoPro Hero5 Session
The action camera is waterproof, up to 10 meters in water, without any need for a special case which is impressive in itself. However, it does not come with any accessories out of the box except for a case and two adhesive mounts. I was expecting more in terms of mounts out of the box.
Features: 7.5/10
The main highlight of the GoPro Hero5 Session is its 4K recording capabilities in such a compact form factor. The 4K recording is at a standard 30 frames per second. The camera comes with an extremely wide angle lens with 10 MP image resolution for photos with effective dimensions of 3648 x 2736 pixels. The shutter button is located on the top along with the menu switching mode at the back of the camera.
GoPro Hero5 Session
You need to use both the buttons to change the modes and cycle through the options given which becomes easier to operate after initial confusion. The camera gives you the option to shoot 4K at frame rates of 30 and 25 frames per second in Wide; 2.7K at frame rates 48, 30, 25, and 24 in SuperView, Wide, Medium and Linear along with the option to shoot 2.7K in 4:3 aspect ratio from the default 16:9.
GoPro Hero5 Session
Other modes include: 2K at 60, 50, 48, 30, 25, and 24 frames per second in Wide mode; 1080p in 90, 60, 50, 48, 30, 25, and 24 frames per second and 960p in 100, 60, 50, 30, and 25 frames per second in Wide mode.
The Camera gives up options to choose modes from Wide, Medium and SuperView while shooting 720p video at 100, 60, 50, 30, and 25 frames per second. You can not go below 720p while shooting with the camera which is weird at best because going down would have given us the option to shoot at higher frame rates for potential use in slow motion videos.
GoPro Hero5 Session
The camera comes with built-in Wi-Fi and Bluetooth for you to connect it to the GoPro ‘Capture’ app for Android as well as an iOS device. You can use the app to record, review or transfer the images and videos that you shot using the camera from the Wi-Fi connection. You can also control all the settings of the action cam and get a live video of what the GoPro can see at any given point.
The camera does support video stabilisation, but that is only limited if you shoot videos up to 2.7K and it is not available if you are shooting 4K. You can use the voice control to tell GoPro to shoot or start recording by saying “GoPro start recording”.
Publish date: April 12, 2017 10:05 am| Modified date: April 12, 2017 10:10 am
Pages: 1 2 3

LeEco announces Le Pro 3 AI Edition smartphone with dual-camera setup in China

LeEco announces Le Pro 3 AI Edition smartphone with dual-camera setup in China

If you remember, there have been numerous leaks around a dual-camera equipped smartphone from LeEco. The wait is finally over. The company has announced the Le Pro 3 AI Edition in China, which is a well-equipped smartphone along with its own AI assistant.
The smartphone has been announced in two variants. There is one powered by a 2.4 GHz MediaTek Helio X23 processor which has 4 GB of RAM and 32 GB of storage. The other one is a more powerful variant running on a Mediatek deca-core 2.6 GHz Helio X27 chipset with 6 GB of RAM and 64 GB of storage.
Rest of the features on both the variants remain the same. There is a 5.5-inch Full HD (1920×1080) display with 600 nits of maximum brightness. There is a dual-camera setup at the rear featuring two 13 MP cameras, one shoots in RGB while the other in monochrome. The cameras have an f/2.0 aperture, PDAF and a dual LED flash. You can also record 4K and 1080p videos at 30 fps as well as slow-mo 720p videos at 120 fps. The front camera is an 8 MP snapper with an f/2.2 aperture.
leeco-le-pro-3-ai-edition (2)
As for the battery, there a large 4,000 mAh unit and the handset runs on EMUI 5.9, which is based on Android Marshmallow. The rest of the features include Wi-Fi 802.11 ac, 4G LTE, Bluetooth 4.2, NFC, an infrared sensor and support for dual SIM.
LeEco also announced its very own AI assistant, which is called, of all things, LeLe. The AI assistant will be integrated with all LeEco apps and will work with a number of third-party applications. One can also activate it from the lockscreen. Just like most smartphone AI assistants, LeLe learns the usage patterns and improves over time.
The 4 GB version of the Le Pro 3 AI Edition is priced 1,799 Yuan (Rs 18,000 approx) while the higher 6 GB model is priced at 2,399 Yuan (Rs 24,000 approx). The smartphone will be available in China from 14 of April and will be available in gold, rose gold, and black colour options.
Publish date: April 12, 2017 11:10 am| Modified date: April 12, 2017 11:10 am

Google’s new machine learning model can harness the processing power of your phone to improve AI

Google’s new machine learning model can harness the processing power of your phone to improve AI

Image: Google
Google’s new machine learning algorithm harnesses the processing power of your smartphone to improve algorithms and save on data consumption by intelligent virtual assistants and the like.
Using a process called Federated Learning (FL), Google claims that they can “collaboratively share a prediction model while keeping training data on the device.” It might sound complicated, but it’s actually a simple, well thought-out solution to a problem that’s been haunting machine learning for years.
Machine Learning (ML) is used to design algorithms. These algorithms help in the functioning of what we know as artificial intelligence (AI). Siri, Google Assistant, Bixby, they’re all just algorithms.
But ML is used for more than just digital assistants. Google’s self-driving car, for example, learns to drive using ML. Essentially, data is harvested and analysed and then an algorithm is developed based on the results. This one-line description grossly oversimplifies how ML works, but you get the general idea.
For example, if you normally drink your coffee at 6 AM every day on weekdays and at 8 AM on weekends, ML will analyse the data and determine your coffee drinking patterns, potentially automating your coffee making process if you have, say, a smart coffee maker. More data might lead the algorithm to determine that you drink coffee late on public holidays as well, or the data might be tied to your heart rate, etc.
The possibilities are endless, but are predicated on a large amount of data, some of it very revealing and potentially a privacy nightmare.
Image: Google
Image: Google
Traditionally, this data would need to be harvested by some device, like your smartphone, and then sent to the cloud for processing. ML is then used to analyse the data further and learn from it, further improving algorithms.
As Google explains in its blog-post, this approach requires low-latency, high-bandwidth data transfer rates. Of course, the privacy issues also raise their head here and taking responsibility for the data can put a company in a tricky spot.
With FL, Google has changed how the learning process works. As they describe it, “your device downloads the current model, improves it by learning from data on your phone, and then summarises the changes as a small focused update.”
The data that’s actually sent to the cloud is the updated algorithm and not the actual data itself. The data remains on your phone at all times.
Google’s cloud-based FL then aggregates these updates and creates an improved model for all devices. This updated model is then sent to devices as an update and the process repeats itself.
Using this technique, Google claims that bandwidth needs can be dropped by up to 100x.
Better yet, for the sake of privacy, the cloud-based processing won’t happen unless “100s or 1000s of users have participated.” This ensures that the updates from a single user cannot be identified.
The updated algorithms that your phone creates are sent to the cloud in a compressed and encrypted manner to begin with.
Google has deployed this technique in their Gboard app to help improve its natural language processing, but as they refine the process, they hope to expand FL’s scope to other applications as well.
FL has the potential to fundamentally change the way models are developed in the future. More information can be found here.
Publish date: April 12, 2017 11:16 am| Modified date: April 12, 2017 11:16 am

Samsung’s voice assistant Bixby could be delayed due to performance issues

Samsung’s voice assistant Bixby could be delayed due to performance issues

With the announcement of its latest flagship duo, the Galaxy S8 and S8+, Samsung also revealed its very own voice assistant ‘Bixby‘. The new AI assistant was touted to work just like Google Assistant. It’s supposed to recommend context-aware information in the form of cards, allow voice control of apps and even allow for control of various smart Samsung devices.
At the time of its announcement, it was said that Bixby will only be available in English and Korean languages, but now there is a report that suggests that the AI assistant is not ready for the former language.
The new flagships are expected to go on sale this month in a variety of markets globally, but it might be missing the new voice assistant. According to The Wall Street Journal, Samsung will not launch the Galaxy S8 and S8+ with Bixby due to certain performance issues discovered in the past few weeks. While the English-language version of Bixby is expected to be delayed until the end of May, a Samsung spokesperson has confirmed that it won’t be available until spring in the US.
Apparently, the performance of the English version of Bixby lagged behind that of the Korean version. Company executives are said to be discussing as to when exactly they want to launch the English version.
This could be a big setback for Samsung as the S8 series has received a good initial response thanks to its amazing design. Then again, if, as we suspect, Bixby won’t be such a big deal, it shouldn’t matter that the S8’s flagship feature is missing.
Publish date: April 12, 2017 11:33 am| Modified date: April 12, 2017 11:33 am

WikiLeaks reveals that NSA has been spying on Pakistan’s mobile networks

WikiLeaks reveals that NSA has been spying on Pakistan’s mobile networks
The US National Security Agency (NSA) has been spying on Pakistan’s mobile networks, whistleblowing organisation WikiLeaks has tweeted. “Hundreds of NSA cyber weapons variants publicly released including code showing hacking of Pakistan mobile system,” @wikileaks tweeted.
According to a report in Express Tribune on Monday, the hacker group “Shadow Brokers” released a new cache of information detailing how the NSA accessed private and public networks in other countries. A researcher on Twitter who identifies himself as ‘x0rz’ decrypted the files and uploaded them on Github, a web-based repository and internet hosting service.
“The researcher confirmed that the archives include evidence of NSA operators’ access inside the GSM network of Mobilink – one of the Pakistan’s most popular mobile services provider,” the report noted. The hacker group had previously released data suggesting the US agency may have been monitoring hundreds of IP addresses in Pakistan. The encrypted files were being decrypted by security researchers around the world.
“Shadow Brokers” had initially wanted to auction its data cache in exchange for Bitcoin but as no buyer turned up, they released the data online. This is not the first time that reports have surfaced claiming that the US NSA is snooping on other countries. According to a Daily Mail report in 2014, WikiLeaks disclosed documents that suggested the Bharatiya Janata Party (BJP) was among six political parties from around the world the NSA was authorised to conduct surveillance on for gathering foreign intelligence. The authorisation was given by a secret American court, it said.
The leak was planned months ahead of Prime Minister Narendra Modi’s visit to the US, the report claimed. Another report in the Washington Post that also came in 2014 said the US has long had broad no-spying arrangements with four countries – Britain, Canada, Australia and New Zealand – in a group known collectively as the “Five Eyes”.
“But a classified 2010 legal certification – approved by the Foreign Intelligence Surveillance Court and included among a set of documents leaked by former NSA contractor Edward Snowden – lists 193 countries [including India], that would be of valid interest for US intelligence,” the Post said.
Publish date: April 12, 2017 11:45 am| Modified date: April 12, 2017 11:41 am

Related Posts Plugin for WordPress, Blogger...