Connect with us


Facebook Researcher’s New Algorithm Ushers New Paradigm Of Image Recognition



“VICReg could be used to model the dependencies between a video clip and the frame that comes after, therefore learning to predict the future in a video.”

Adrien Bardes, Facebook AI Research

Humans have an innate capability to identify objects in the wild, even from a blurred glimpse of the thing. We do this efficiently by remembering only high-level features that get the job done (identification) and ignoring the details unless required. In the context of deep learning algorithms that do object detection, contrastive learning explored the premise of representation learning to obtain a large picture instead of doing the heavy lifting by devouring pixel-level details. But, contrastive learning has its own limitations. 

According to Andrew Ng, pre-training methods can suffer from three common failings: generating an identical representation for different input examples (which leads to predicting the mean consistently in linear regression), generating dissimilar representations for examples that humans find similar (for instance, the same object viewed from two angles), and generating redundant parts of a representation (say, multiple vectors that represent two eyes in a photo of a face). The problems of representation learning, wrote Andrew Ng, boil down to variance, invariance, and covariance issues. 

Register for our upcoming AI Conference>>

Also Read: What is Contrastive Learning

Andrew Ng’s observations are a reference to a new self-supervised algorithm released by the researchers at Facebook AI, PSL Research University, and New York University, along with Turing award recipient Yann Lecun introduced called Variance-Invariance-Covariance Regularization (VICReg), which builds on Lecun’s own Barlow Twins method.

free widgets for website
(Image credits: Bardes et al.,)

The researchers designed VICReg (Variance-Invariance-Covariance Regularization) to avoid the collapse problem, which is handled more inefficiently in the case of contrastive methods. They do this by introducing a simple regularisation term on the variance of the embeddings along each dimension individually and combining the variance term with a decorrelation mechanism based on redundancy reduction and covariance regularisation. The authors state that VICReg is performed on par with several state-of-the-art methods.

VICReg is a simple approach to self-supervised image representation learning, and its objectives are as follows:

  • Learn invariance to different views with an invariance term.
  • Avoid collapse of the representations with a variance regularisation term.
  • Spread the information throughout the different dimensions of the representations with a covariance regularisation term. 

The results show that VICReg performs on par with state-of-the-art methods and ushers a new paradigm of non-contrastive self-supervised learning. 

What Authors Had To Say

VICReg: Variance-Invariance-Covariance Regularization for Self-Supervised Learning.

By Adrien Bardes, Jean Ponce, and yours truly.

Insanely simple and effective method for self-supervised training of joint-embedding architectures (e.g. Siamese nets).


— Yann LeCun (@ylecun) May 12, 2021

Talking to Analytics India Magazine about VICReg’s significance, the lead author, Adrien Bardes, who is also a resident PhD student at Facebook AI Research, Paris, said that self-supervised representation learning is a learning paradigm that aims to learn meaningful representations of some unlabelled data. Recent approaches rely on Siamese networks and maximise the similarity between two augmented views of the same input. A trivial solution is for the network to output constant vectors, known as the collapse problem. VICReg is a new algorithm based on siamese networks but aims to prevent a collapse by regularising the variance and covariance of the network outputs. It achieves state-of-the-art results in several computer vision benchmarks while being a straightforward and interpretable approach.

When asked about how VICReg addresses shortcomings of contrastive learning methods, Bardes explained that contrastive learning methods are based on a simple principle. They make the inputs that should encode similar information close to each other in the embedding space and prevent a collapse by pushing apart the inputs that should encode dissimilar information. This process requires the mining of a massive amount of negative pairs, pairs of distinct inputs. Recent contrastive approaches for self-supervised learning have different strategies for mining these negative pairs; they can sample them from a memory bank, as in MoCo, or sample them from the current batch, as in SimCLR, which in both cases is costly in time or memory. VICReg, on the other hand, does not require these negative pairs; it implicitly prevents a collapse by enforcing the representations to be different from each other without making any direct comparison between different examples. It, therefore, does not require the memory bank of MoCo and works with much smaller batch sizes than SimCLR.

For Bardes, self-supervised learning is probably the most exciting topic in machine learning research. Annotating data is a very expansive process performed by humans who have biases and can make mistakes. It is, therefore, impossible to annotate the vast amount of data available today, for example, medical or astronomical data and images and videos on the Internet. Training models that leverage all these data can only be done using self-supervised learning. This is one of the motivations behind the development of VICReg.

free widgets for website

Bardes believes that VICReg is applicable in any scenario where one wants to model the relationships within a data set. It can be used with any kind of data, images, videos, text, audio, or proteins. For example, you could use it to model the dependencies between a video clip and the frame after, therefore learning to predict the future in a video. Another example would be to understand the relationship between the graph of a molecule and its image seen from a microscope.

“We are at the early stages of the development of self-supervised learning. Shifting from contrastive methods to non-contrastive methods is the first step towards more practical algorithms. Current approaches rely on hand-craft data augmentations that can be viewed as a kind of supervision. The next step will probably be to get rid of these augmentations. Another promising direction consists in handling the uncertainties in modelling the data. Current methods are mostly deterministic and always model the same relation between two inputs. For example, if we go back to the frame prediction example, current methods would only model the possible future for a video clip. Future approaches will probably use latent variables that model the space of possible predictions,” concluded Bardes.

Join Our Discord Server. Be part of an engaging online community. Join Here.

Subscribe to our Newsletter

Get the latest updates and relevant offers by sharing your email.

Read More

free widgets for website
See also  Digital Reality Series Collab Crib Debuts on "Facebook Watch" | |


Updating Special Ad Audiences for housing, employment, and credit advertisers





On June 21, 2022 we announced an important settlement with the US Department of Housing and Urban Development (HUD) that will change the way we deliver housing ads to people residing in the US. Specifically, we are building into our ads system a method designed to make sure the audience that ends up seeing a housing ad more closely reflects the eligible targeted audience for that ad.

As part of this agreement, we will also be sunsetting Special Ad Audiences, a tool that lets advertisers expand their audiences for ad sets related to housing. We are choosing to sunset this for employment and credit ads as well. In 2019, in addition to eliminating certain targeting options for housing, employment and credit ads, we introduced Special Ad Audiences as an alternative to Lookalike Audiences. But the field of fairness in machine learning is a dynamic and evolving one, and Special Ad Audiences was an early way to address concerns. Now, our focus will move to new approaches to improve fairness, including the method previously announced.

What’s happening: We’re removing the ability to create Special Ad Audiences via Ads Manager beginning on August 25, 2022.

Beginning October 12th, 2022, we will pause any remaining ad sets that contain Special Ad Audiences. These ad sets may be restarted once advertisers have removed any and all Special Ad Audiences from those ad sets. We are providing a two month window between preventing new Special Ad Audiences and pausing existing Special Ad Audiences to enable advertisers the time to adjust budgets and strategies as needed.

See also  Zuckerberg outlines how Facebook will thrive after Apple privacy change

For more details, please visit our Newsroom post.

free widgets for website

Impact to Advertisers using Marketing API on September 13, 2022

For advertisers and partners using the API listed below, the blocking of new Special Ad Audience creation will present a breaking change on all versions. Beginning August 15, 2022, developers can start to implement the code changes, and will have until September 13, 2022, when the non-versioning change occurs and prior values are deprecated. Refer below to the list of impacted endpoints related to this deprecation:

For reading audience:

  • endpoint gr:get:AdAccount/customaudiences
  • field operation_status

For adset creation:

  • endpoint gr:post:AdAccount/adsets
  • field subtype

For adset editing:

  • endpoint gr:post:AdCampaign
  • field subtype

For custom audience creation:

  • endpoint gr:post:AdAccount/customaudiences
  • field subtype

For custom audience editing:

  • endpoint gr:post:CustomAudience

Please refer to the developer documentation for further details to support code implementation.

First seen at

free widgets for website
Continue Reading


Introducing an Update to the Data Protection Assessment





Over the coming year, some apps with access to certain types of user data on our platforms will be required to complete the annual Data Protection Assessment. We have made a number of improvements to this process since our launch last year, when we introduced our first iteration of the assessment.

The updated Data Protection Assessment will include a new developer experience that is enhanced through streamlined communications, direct support, and clear status updates. Today, we’re sharing what you can expect from these new updates and how you can best prepare for completing this important privacy requirement if your app is within scope.

If your app is in scope for the Data Protection Assessment, and you’re an app admin, you’ll receive an email and a message in your app’s Alert Inbox when it’s time to complete the annual assessment. You and your team of experts will then have 60 calendar days to complete the assessment. We’ve built a new platform that enhances the user experience of completing the Data Protection Assessment. These updates to the platform are based on learnings over the past year from our partnership with the developer community. When completing the assessment, you can expect:

  • Streamlined communication: All communications and required actions will be through the My Apps page. You’ll be notified of pending communications requiring your response via your Alerts Inbox, email, and notifications in the My Apps page.

    Note: Other programs may still communicate with you through the App Contact Email.

  • Available support: Ability to engage with Meta teams via the Support tool to seek clarification on the questions within the Data Protection Assessment prior to submission and help with any requests for more info, or to resolve violations.

    Note: To access this feature, you will need to add the app and app admins to your Business Manager. Please refer to those links for step-by-step guides.

  • Clear status updates: Easy to understand status and timeline indicators throughout the process in the App Dashboard, App Settings, and My Apps page.
  • Straightforward reviewer follow-ups: Streamlined experience for any follow-ups from our reviewers, all via

We’ve included a brief video that provides a walkthrough of the experience you’ll have with the Data Protection Assessment:

Something Went Wrong

free widgets for website

We’re having trouble playing this video.

The Data Protection Assessment elevates the importance of data security and helps gain the trust of the billions of people who use our products and services around the world. That’s why we are committed to providing a seamless experience for our partners as you complete this important privacy requirement.

Here is what you can do now to prepare for the assessment:

  1. Make sure you are reachable: Update your developer or business account contact email and notification settings.
  2. Review the questions in the Data Protection Assessment and engage with your teams on how best to answer these questions. You may have to enlist the help of your legal and information security points of contact to answer some parts of the assessment.
  3. Review Meta Platform Terms and our Developer Policies.

We know that when people choose to share their data, we’re able to work with the developer community to safely deliver rich and relevant experiences that create value for people and businesses. It’s a privilege we share when people grant us access to their data, and it’s imperative that we protect that data in order to maintain and build upon their trust. This is why the Data Protection Assessment focuses on data use, data sharing and data security.

Data privacy is challenging and complex, and we’re dedicated to continuously improving the processes to safeguard user privacy on our platform. Thank you for partnering with us as we continue to build a safer, more sustainable platform.

First seen at

free widgets for website
See also  Facebook content moderator who quit reportedly wrote a blistering letter citing 'stress induced ...
Continue Reading


Resources for Completing App Store Data Practice Questionnaires for Apps That Include the Facebook or Audience Network SDK





Updated July 18: Developers and advertising partners may be required to share information on their app’s privacy practices in third party app stores, such as Google Play and the Apple App Store, including the functionality of SDKs provided by Meta. To help make it easier for you to complete these requirements, we have consolidated information that explains our data collection practices for the Facebook and Audience Network SDKs.

Facebook SDK

To provide functionality within the Facebook SDK, we may receive and process certain contact, location, identifier, and device information associated with Facebook users and their use of your application. The information we receive depends on what SDK features 3rd party applications use and we have structured the document below according to these features.

App Ads, Facebook Analytics, & App Events

Facebook App Events allow you to measure the performance of your app using Facebook Analytics, measure conversions associated with Facebook ads, and build audiences to acquire new users as well as re-engage existing users. There are a number of different ways your app can use app events to keep track of when people take specific actions such as installing your app or completing a purchase.

With Facebook SDK, there are app events that are automatically logged (app installs, app launches, and in-app purchases) and collected for Facebook Analytics unless you disable automatic event logging. Developers determine what events to send to Facebook from a list of standard events, or via a custom event.

When developers send Facebook custom events, these events could include data types outside of standard events. Developers control sending these events to Facebook either directly via application code or in Events Manager for codeless app events. Developers can review their code and Events Manager to determine which data types they are sending to Facebook. It’s the developer’s responsibility to ensure this is reflected in their application’s privacy policy.

free widgets for website

Advanced Matching

Developers may also send us additional user contact information in code, or via the Events Manager. Advanced matching functionality may use the following data, if sent:

  • email address, name, phone number, physical address (city, state or province, zip or postal code and country), gender, and date of birth.
See also  EXPLAINER-Google, Facebook battle Australia over proposed revenue-sharing law

Facebook Login

There are two scenarios for applications that use Facebook Login via the Facebook SDK: Authenticated Sign Up or Sign In, and User Data Access via Permissions. For authentication, a unique, app-specific identifier tied to a user’s Facebook Account enables the user to sign in to your app. For Data Access, a user must explicitly grant your app permission to access data.

Note: Since Facebook Login is part of the Facebook SDK, we may collect other information referenced here when you use Facebook Login, depending on your settings.

Device Information

We may also receive and process the following information if your app is integrated with the Facebook SDK:

  • Device identifiers;
  • Device attributes, such as device model and screen dimensions, CPU core, storage size, SDK version, OS and app versions, and app package name; and
  • Networking information, such as the name of the mobile operator or ISP, language, time zone, and IP address.

Audience Network SDK

We may receive and process the following information when you use the Audience Network SDK to integrate Audience Network ads in your app:

  • Device identifiers;
  • Device attributes, such as device model and screen dimensions, operating system, mediation platform and SDK versions; and
  • Ad performance information, such as impressions, clicks, placement, and viewability.

First seen at

Continue Reading