Connect with us
Machine learning can help Twitch catch banned violators Machine learning can help Twitch catch banned violators

Security

Machine learning can help Twitch catch banned violators

Published

on

The company is activating a tool on its platform to catch users trying to bypass the “ban”.

Twitch continues to fight harassment on its platform. She recently armed herself with a new tool that uses machine learning to identify users trying to bypass a ban.

A new tool called Suspicious User Detection identifies users as “likely” or “possible” intruders trying to bypass a ban. The underlying machine learning model identifies potential bias by comparing their behavior and account characteristics with information from locked out accounts.

Messages from users flagged as “likely” offendants will not appear in Twitch chat, but streamers and moderators will be able to see them. Streamers and moderators will be able to choose for themselves whether to monitor the “likely” biaser (in this case, the user will be added to the monitoring list, and his messages will be marked as subject to monitoring) or block him. Messages from “possible” dodgers will be displayed in the chat, but streamers and moderators will be able to block them at their discretion.

According to Twitch, the Suspicious User Detection tool will be enabled by default. Streamers and moderators will also be able to manually select users to monitor whose behavior they find suspicious.

“The only thing you need to be prepared for, especially immediately after launch (Suspicious User Detection – ed.), Is that no machine learning model can be 100% accurate, which means possible false positive and false negative results. Therefore, Suspicious User Detection itself does not “ban” possible and likely violators. When it comes to the community, you are the experts, and you must make the final decision who can participate. The tool will learn from the actions you take, and as a result, the accuracy of its predictions should increase over time “, – reported on the Twitch blog.

Click to comment

Leave a Reply

Your email address will not be published.

Security

Nvidia lost. LHR mining protection is also hacked under Linux. This was done by NBMiner developers

Published

on

Nvidia lost LHR mining protection is also hacked under

Two days have passed since the NiceHash developers cracked the Nvidia LHR protection, as the NBMiner team also pleased their users with the same news. Only this time we are talking about software for Linux.

Nvidia lost.  LHR mining protection is also hacked under Linux.  This was done by NBMiner developers

Thus, Nvidia’s protection completely fell for both Windows and Linux. Unfortunately, both programs are closed source, so it’s not clear what mechanisms the developers used to hack.

Whether the loss of Nvidia will affect the availability and prices of video cards is still difficult to say. At the moment, the cryptocurrency market continues to fall, but sooner or later it will turn around, and gamers may again face shortages and overpriced graphics cards.

Continue Reading

Security

Xiaomi has released a profitable set of security camera and smart door locksmart door lock

Published

on

Xiaomi has released a profitable set of security camera and

Xiaomi has introduced a new kit with an outdoor video surveillance camera and a smart door lock, which includes Mi Outdoor camera and Mi Smart Door Lock 1S.

The bundle is priced at around $237, which is a great deal as these devices cost $20 more individually.

Mi Smart Door Lock 1S supports 7 unlocking methods, including fingerprint, password, temporary password, Bluetooth, HomeKit, NFC or regular key unlock. Compared to the first generation, the new lock supports both the Mijia app and Apple HomeKit.

Xiaomi has released a profitable set of security camera and

As for the rechargeable version of Xiaomi Outdoor Camera, this is Xiaomi’s first outdoor wireless camera. It has an independent design and can be installed without connecting the mains cable or power cable. It has a wide viewing angle of 130°, 1080p resolution and supports WDR technology.

In addition, the battery version of the Xiaomi Outdoor Camera offers night vision up to 7 meters and people detection function. It is IP65 rated and has a long battery life of up to 90 days.

Continue Reading

Security

“These are machines for sucking out personal data.” Prayer and mental health apps have poor security

Published

on

These are machines for sucking out personal data Prayer and

Mental health apps have worse privacy protection than most other types of apps, according to a new analysis by Mozilla. We are talking about the entire category as a whole. In addition, things are also bad for prayer applications.

“These are machines for sucking out personal data.”  Prayer and mental health apps have poor security

The vast majority of mental health and prayer apps are exceptionally creepy. They track a variety of data, share and capitalize on users’ most intimate personal thoughts and feelings, such as mood, mental state, and biometric data.

The team analyzed 32 mental health and prayer apps. Of these apps, 29 received a Privacy not included warning, indicating that the team is concerned about how the app manages user data.

These applications collect large amounts of personal data in accordance with vague privacy policies. Most applications have also been found to have poor security practices that allow users to create accounts with weak passwords. Considering how much personal information such programs can contain, this is a bad feature.

The list of the worst programs according to the specified criteria included Better Help, Youper, Woebot, Better Stop Suicide, Pray.com and Talkspace. In particular, the Woebot chatbot claims to collect information about users and shares this data for advertising purposes, and Talkspace collects transcripts of user chats.

They work like data-sucking machines with the look and feel of a mental health app. In other words: wolves in sheep’s clothing

Continue Reading

Most Popular