Connect with us

FACEBOOK

Computer vision inches toward ‘common sense’ with Facebook’s latest research

Published

on

Machine learning is capable of doing all sorts of things as long as you have the data to teach it how. That’s not always easy, and researchers are always looking for a way to add a bit of “common sense” to AI so you don’t have to show it 500 pictures of a cat before it gets it. Facebook’s newest research takes a big step toward reducing the data bottleneck.

The company’s formidable AI research division has been working for years now on how to advance and scale things like advanced computer vision algorithms, and has made steady progress, generally shared with the rest of the research community. One interesting development Facebook has pursued in particular is what’s called “semi-supervised learning.”

Generally when you think of training an AI, you think of something like the aforementioned 500 pictures of cats — images that have been selected and labeled (which can mean outlining the cat, putting a box around the cat or just saying there’s a cat in there somewhere) so that the machine learning system can put together an algorithm to automate the process of cat recognition. Naturally if you want to do dogs or horses, you need 500 dog pictures, 500 horse pictures, etc. — it scales linearly, which is a word you never want to see in tech.

Semi-supervised learning, related to “unsupervised” learning, involves figuring out important parts of a data set without any labeled data at all. It doesn’t just go wild, there’s still structure; for instance, imagine you give the system a thousand sentences to study, then showed it 10 more that have several of the words missing. The system could probably do a decent job filling in the blanks just based on what it’s seen in the previous thousand. But that’s not so easy to do with images and video — they aren’t as straightforward or predictable.

But Facebook researchers have shown that while it may not be easy, it’s possible and in fact very effective. The DINO system (which stands rather unconvincingly for “DIstillation of knowledge with NO labels”) is capable of learning to find objects of interest in videos of people, animals and objects quite well without any labeled data whatsoever.

Animation showing four videos and the AI interpretation of the objects in them.

Image Credits: Facebook

It does this by considering the video not as a sequence of images to be analyzed one by one in order, but as a complex, interrelated set, like the difference between “a series of words” and “a sentence.” By attending to the middle and the end of the video as well as the beginning, the agent can get a sense of things like “an object with this general shape goes from left to right.” That information feeds into other knowledge, like when an object on the right overlaps with the first one, the system knows they’re not the same thing, just touching in those frames. And that knowledge in turn can be applied to other situations. In other words, it develops a basic sense of visual meaning, and does so with remarkably little training on new objects.

This results in a computer vision system that’s not only effective — it performs well compared with traditionally trained systems — but more relatable and explainable. For instance, while an AI that has been trained with 500 dog pictures and 500 cat pictures will recognize both, it won’t really have any idea that they’re similar in any way. But DINO — although it couldn’t be specific — gets that they’re similar visually to one another, more so anyway than they are to cars, and that metadata and context is visible in its memory. Dogs and cats are “closer” in its sort of digital cognitive space than dogs and mountains. You can see those concepts as little blobs here — see how those of a type stick together:

Animated diagram showing how concepts in the machine learning model stay close together.

Image Credits: Facebook

This has its own benefits, of a technical sort we won’t get into here. If you’re curious, there’s more detail in the papers linked in Facebook’s blog post.

There’s also an adjacent research project, a training method called PAWS, which further reduces the need for labeled data. PAWS combines some of the ideas of semi-supervised learning with the more traditional supervised method, essentially giving the training a boost by letting it learn from both the labeled and unlabeled data.

Facebook of course needs good and fast image analysis for its many user-facing (and secret) image-related products, but these general advances to the computer vision world will no doubt be welcomed by the developer community for other purposes.

TechCrunch

FACEBOOK

5 Effective Ways to Run Facebook Ads A/B Tests

Published

on

Testing-and-Scaling-Facebook-Ads

Facebook Ads A/B Tests or split tests help them try different versions of ads with various campaign elements. This process helps them arrive at the best version for the organization’s target. 

A/B Tests offer a vast pool of resources to try out various versions. You may get caught up and lose your way to arriving at the best version in a limited time. To better understand this topic you can read the Facebook ad testing guide. Here are five effective ways to run Facebook Ads A/B Tests-

1) Start with the minimal number of variables

This approach will help you analyze the impact of a variable much better. The lesser the variables, the better will be the relevant results and more conclusive. Once you have various versions, you will need to run them through the A/B Significance Test to determine if the test results are valid.

2) The second way is to select the correct structure. 

There are two structures in A/B tests. One is a single ad test, and the other is multiple single variation ad sets. All the variations will go under one ad set in the first structure. Each variation will be under a separate ad set in the second one. Out of the two, the second one works out to be better and gives better results.

3) Use of spreadsheets is important to stay organized. 

These spreadsheets help collect and analyze data to get meaningful insights and arrive at data-backed decisions.

4) Do target advertising and set realistic time goals. 

One approach is to choose an entirely new set of audiences. Also, the data pool should be vast and not the same as some existing campaigns. The reason for choosing a different audience is that Facebook may mix up your ads and give contaminated output. 

Another approach to choosing the right audience is to pick geography. It works better, especially when you have business in a particular region.   

It’s also essential to set a realistic timeline for your testing. Facebook suggests one should run a test for at least four days, but you can choose to run the test for up to 30 days.   

5) Set an ideal budget. 

The concept of a perfect budget is subjective. But, you can fix it yourself, or Facebook can do that for you based on your testing data. A large part of the test budget is spent on avoiding audience duplication. If the same audience sees variations, it could affect the test results.

Besides these top five effective ideas, you will need to take a few more action points to make the testing process efficient. Make sure you put the website’s domain link and not the landing page link in the ad, as that doesn’t look good. Put appropriate Call To Action Button, such as ‘Learn More,’ ‘Buy Now,’ etc. It’s also important to see how your ad is coming across on various electronic gadgets- mobile, tablets, etc.

Another strategy that works is trying to engage the customer. You may add social engagement buttons such as ‘Like’ or ‘Comment.’ Use high-resolution images as they work better with the customers. Low-quality, highly edited images are often not liked and trusted by the consumers.

You can learn more about the audience behavior patterns with A/B test results. Conducting these tests on Facebook streamlines the entire process and makes it smooth for you. With the test results, advertisers and marketers can work on the creatives they need to utilize.

To sum it up, you can run an effective A/B test campaign within the specified budget. You don’t need to spend massive amounts to get your advertisement right. You’ll make the correct assumptions about the performance of variations with a good understanding of business and consumers.

Source link

Continue Reading

DON'T MISS ANY IMPORTANT NEWS!
Subscribe To our Newsletter
We promise not to spam you. Unsubscribe at any time.
Invalid email address

Trending

en_USEnglish