Connect with us


Meta Announces Research To Create Human-Level AI



Meta Announces Research To Create Human-Level AI

Facebook parent Meta announced they are launching a long-term research project to build a next-generation AI that can learn and process speech and text in the same way the human brain does. Meta described an effort to create a human-level AI.

Meta is partnering with a neuroimaging company, NeuroSpin, which images the human brain and with a software company Inria, to study how the human brain processes speech and text and then compare that with how AI language models.

NeuroSpin is a research center that is specifically focused on brain imaging. The researchers consist of physicists, mathematicians, neuroscientists and doctors who work together to build tools to learn about the human brain in different ways.

NeuroSpin explains what it does:

“Focused on neuroimaging, the research conducted ranges from technological and methodological developments (data acquisition and processing) to preclinical and clinical neuroscience, including cognitive neuroscience.”

Meta published:

“Today, we’re announcing a long-term AI research initiative to better understand how the human brain processes speech and text. In collaboration with neuroimaging center Neurospin (CEA) and Inria, we’re comparing how AI language models and the brain respond to the same spoken or written sentences.

We’ll use insights from this work to guide the development of AI that processes speech and text as efficiently as people.”

The problem with AI language models is that they need a lot of examples in order to learn. Human brains need only a few examples to learn.

The current research into brain-like AI language models discovered:

“Language models that most resemble brain activity are those that best predict the next word from context (e.g. once upon a …time).

While the brain anticipates words and ideas far ahead in time, most language models are trained to only predict the very next word. Unlocking this long-range forecasting capability could help improve modern AI language models.”

The announcement cited current research into modeling AI on human brain activity that used MRIs and other imaging tools to view human brain activity when the humans were accomplishing various language-related tasks.

The research paper cited is from 2021 and it is titled, Language processing in brains and deep neural networks: computational convergence and its limits (PDF).

A summary of the findings is discussed in the opening paragraphs of the research paper:

“The results show that (1) the position of the layer in the network and (2) the ability of the network to accurately predict words from context are the main factors responsible for the emergence of brain-like representations in artificial neural networks.

Together, these results show how perceptual, lexical and compositional representations precisely unfold within each cortical region and contribute to uncovering the governing principles of language processing in brains and algorithms.”

The importance of the above research is to show how researching how the brain processes data can yield insights into creating similar processes in an algorithm.

The Meta research teams are using thousands of scans of human brain activity to see which regions of the brain were activated during tasks.

This research was said to show the “computational organization of the human brain” which yielded insights useful toward Meta’s goal of developing “human-level AI.”

The benefits aren’t limited to generating human-level AI, the research also helps neuroscientists better understand the human brain.


Read the Official Meta Announcement

Building AI That Processes Language as People Do

Read a More In-Depth Description of Meta’s Human-Level AI Research

Studying the brain to build AI that processes language as people do

Source link

Keep an eye on what we are doing
Be the first to get latest updates and exclusive content straight to your email inbox.
We promise not to spam you. You can unsubscribe at any time.
Invalid email address


Site Quality Is Simpler Than People Think




Site Quality Is Simpler Than People Think

Google’s John Mueller, Martin Splitt and Gary Illyes discussed site quality in a recent podcast, explaining the different ways of thinking about site quality and at one point saying it’s not rocket science. The discussion suggests that site quality could be simpler than most people know.

Site Quality Is Not Rocket Science

The first point they touched on is to recommend reading site quality documentation, insisting that site quality is not especially difficult to understand.

Gary Illyes said:

“So I would go to a search engine’s documentation.

Most of them have some documentation about how they function and just try to figure out where your content might be failing or where your page might be failing because honestly, okay, this is patronizing, but it’s not rocket science.”

No Tools For Site Quality – What To Do?

Gary acknowledged that there’s no tool for diagnosing site quality, not in the same way there are tools for objectively detecting technical issues.

The traffic metrics that show a downward movement don’t explain why, they just show that something changed.

Gary Illyes:

“I found the up-down metric completely useless because you still have to figure out what’s wrong with it or why people didn’t like it.

And then you’re like, “This is a perfectly good page. I wrote it, I know that it’s perfect.”

And then people, or I don’t know, like 99.7% of people are downvoting it. And you’re like, ‘Why?’”

Martin Splitt

“And I think that’s another thing.

How do I spot, I wrote the page, so clearly it is perfect and helpful and useful and amazing, but then people disagree, as you say.

How do you think about that? What do you do then?

How can I make my content more helpful, better, more useful? I don’t know.

…There’s all these tools that I can just look at and I see that something’s good or something’s bad.

But for quality, how do I go about that?”

Gary Illyes

“What if quality is actually simpler than at least most people think?

…What if it’s about writing the thing that will help people achieve whatever they need to achieve when they come to the page? And that’s it.”

Martin Splitt asked if Gary was talking about reviewing the page from the perspective of the user.

Illyes answered:

“No, we are reframing.”

Reframing generally means to think about the problem differently.

Gary’s example is to reframe the problem as whether the page delivers what it says it’s going to deliver (like helping users achieve X,Y,Z).

Something I see a lot with content is that the topic being targeted (for example, queries about how to catch a trout) isn’t matched by the content (which might actually be about tools for catching trout) which is not what the site visitor wants to achieve.

Quality In Terms Of Adding Value

There are different kinds of things that relate to site and page quality and in the next part of the podcast John Mueller and Gary Illyes discuss the issue about adding something of value.

Adding something of value came up in the context of where the SERPs offer good answers from websites that people not only enjoy but they expect to see those sites as answers for those queries.

You can tell when users expect specific sites for individual search queries when Google Suggests shows the brand name and the keyword.

That’s a clue that probably a lot of people are turning keywords into branded searches, which signals to Google what people want to see.

So, the problem of quality in those situations isn’t about being relevant for a query with the perfect answer.

For these situations, like for competitive queries, it’s not enough to be relevant or have the perfect answer.

John Mueller explains:

“The one thing I sometimes run into when talking with people is that they’ll be like, “Well, I feel I need to make this page.”

And I made this page for users in air quotes…

But then when I look at the search results, it’s like 9,000 other people also made this page.

It’s like, is this really adding value to the Internet?

And that’s sometimes kind of a weird discussion to have.

It’s like, ‘Well, it’s a good page, but who needs it?’

There are so many other versions of this page already, and people are happy with those.”

This is the type of situation where competitive analysis to “reverse engineer” the SERPs  works against the SEO.

It’s stale because using what’s in the SERPs as a template for what to do rank is feeding Google what it already has.

It’s like, as an example, let’s represent the site ranked in Google with a baseline of the number zero.

Let’s imagine everything in the SERPs has a baseline of zero. Less than zero is poor quality. Higher than zero is higher quality.

Zero is not better than zero, it’s just zero.

The SEOs who think they’re reverse engineering Google by copying entities, copying topics, they’re really just achieving an imperfect score of zero.

So, according to Mueller, Google responds with, “it’s a good page, but who needs it?”

What Google is looking for in this situation is not the baseline of what’s already in the SERPs, zero.

According to Mueller, they’re looking for something that’s not the same as the baseline.

So in my analogy, Google is looking for something above the baseline of what is already in the SERPs, a number greater than zero, which is a one.

You can’t add value by feeding Google back what’s already there. And you can’t add value by doing the same thing ten times bigger. It’s still the same thing.

Breaking Into The SERPs By The Side Door

Gary Illyes next discusses a way to break into a tough SERP, saying the way to do it is indirectly.

This is an old strategy but a good one that still works today.

So, rather than bringing a knife to a gunfight, Gary Illyes suggests choosing more realistic battles to compete in.

Gary continued the conversation about competing in tough SERPs.

He said:

“…this also is kind of related to the age-old topic that if you are a new site, then how can you break into your niche?

I think on today’s Internet, like back when I was doing ‘SEO’, it was already hard.

For certain topics or niches, it was absolutely a nightmare, like ….mesothelioma….

That was just impossible to break into. Legal topics, it was impossible to break into.

And I think by now, we have so much content on the Internet that there’s a very large number of topics where it is like 15 years ago or 20 years ago, that mesothelioma topic, where it was impossible to break into.

…I remember Matt Cutts, former head of Web Spam, …he was doing these videos.

And in one of the videos, he said try to offer something unique or your own perspective to the thing that you are writing about.

Then the number of perspective or available perspectives, free perspectives, is probably already gone.

But if you find a niche where people are not talking too much about, then suddenly, it’s much easier to break into.

So basically, this is me saying that you can break into most niches if you know what you are doing and if you are actually trying to help people.”

What Illyes is suggesting as a direction is to “know what you are doing and if you are actually trying to help people.

That’s one of my secrets to staying one step ahead in SEO.

For example, before the reviews update, before Google added Experience to E-A-T, I was telling clients privately to do that for their review pages and I told them to keep it a secret, because I knew I had it dialed in.

I’m not psychic, I was just looking at what Google wants to rank and I figured it out several years before the reviews update that you need to have original photos, you need to have hands-on experience with the reviewed product, etc.

Gary’s right when he advises to look at the problem from the perspective of “trying to help people.”

He next followed up with this idea about choosing which battles to fight.

He said:

“…and I think the other big motivator is, as always, money. People are trying to break into niches that make the most money. I mean, duh, I would do the same thing probably.

But if you write about these topics that most people don’t write about, let’s say just three people wrote about it on the Internet, then maybe you can capture some traffic.

And then if you have many of those, then maybe you can even outdo those high-traffic niches.”

Barriers To Entry

What Gary is talking about is how to get around the barrier to entry, which are the established sites. His suggestion is to stay away from offering what everyone else is offering (which is a quality thing).

Creating content that the bigger sites can’t or don’t know to create is an approach I’ve used with a new site.

Weaknesses can be things that the big site does poorly, like their inability to resonate with a younger or older audience and so on.

Those are examples of offering something different that makes the site stand out from a quality perspective.

Gary is talking about picking the battles that can be won, planting a flag, then moving on to the next hill.

That’s a far better strategies than walking up toe to toe with the bigger opponent.

Analyzing For Quality Issues

It’s a lot easier to analyze a site for technical issues than it is for quality issues.

But a few of the takeaways are:

  • Be aware that the people closest to the content are not always the best judges of content is quality.
  • Read Google’s search documentation (for on-page factors, content, and quality guidelines).
  • Content quality is simpler than it seems. Just think about knowing the topic well and being helpful to people.
  • Being original is about looking at the SERPs for things that you can do differently, not about copying what the competitors are doing.

In my experience, it’s super important to keep an open mind, to not get locked into one way of thinking, especially when it comes to site quality. This will help one keep from getting locked into a point of view that can keep one from seeing the true cause of ranking issues.

Featured Image by Shutterstock/Stone36

Source link

Keep an eye on what we are doing
Be the first to get latest updates and exclusive content straight to your email inbox.
We promise not to spam you. You can unsubscribe at any time.
Invalid email address
Continue Reading


Is Alt Text A Ranking Factor For Google Image Search?




Is Alt Text A Ranking Factor For Google Image Search?

Alt text is used to help computers read images.

But can alt tags affect your organic search rankings?

Read on to learn whether there is any connection between alt text and improved rankings in Google Image Search results.

The Claim: Alt Text Is A Ranking Factor

What is alt text?

Alt text is an HTML image attribute. It allows you to create an alternative text version of your image if it cannot load or has an accessibility issue.

Because of its importance to Google Image Search, it is considered a ranking factor.

[Ranking Factors 2023] Download the free ebook + cheat sheet 

Alt Text As A Ranking Factor: The Evidence

Google emphasizes how alt text plays a vital role in getting your images recognized by Google Image Search.

You will find a page on image best practices in Google Search Central’s Advanced SEO documentation. In a section called “about alt text,” Google discusses the use of alt text.

“Google uses alt text along with computer vision algorithms and the contents of the page to understand the subject matter of the image. Also, alt text in images is useful as anchor text if you decide to use an image as a link.”

While the company doesn’t specify that alt text will improve your rankings, it warns website owners that improper use can harm your website.

“When writing alt text, focus on creating useful, information-rich content that uses keywords appropriately and is in context of the content of the page.

Avoid filling alt attributes with keywords (also known as keyword stuffing) as it results in a negative user experience and may cause your site to be seen as spam.”

It also offers the following examples of good and bad alt text usage.

Screenshot from Google Search Central, August 2023Google Search Central best practice for images

Google Sites Help documentation indicates that images may come with pre-populated alt text, including keywords for which you may not want to optimize.

“Some images automatically include alt text, so it’s a good idea to check that the alt text is what you want.”

For example, when I download stock photos, a text description of the image is embedded in the file.

Is Alt Text A Ranking Factor For Google Image Search?Screenshot by author, August 2023Is Alt Text A Ranking Factor For Google Image Search?

When uploaded to a content management system (CMS) like WordPress, the text descriptions may need to be moved to the alt text field or modified to remove unnecessary keywords.

Is Alt Text A Ranking Factor For Google Image Search?Screenshot from WordPress, August 2023Is Alt Text A Ranking Factor For Google Image Search?

In Google Search Central’s “Search Engine Optimization Starter Guide,” it offers the following advice about alt tags when using images as links:

“…if you’re using an image as a link, the alt text for that image will be treated similarly to the anchor text of a text link. However, we don’t recommend using too many images for links in your site’s navigation when text links could serve the same purpose.”

In 2020, John Mueller, Google Search Advocate, answered a question about the alt text of a quote image during a Google Webmaster Office Hours. In the answer, he talked about how Google uses it:

“For Search, what happens with the alt attribute is we use that to better understand the images themselves, in particular, for Image Search. So if you didn’t care about Image Search, then from a Search point of view, you don’t really need to worry about alt text.

But if you do want these images to be shown in Image Search, which sometimes it makes sense to show fancy quotes in Image Search as well, then using the alt attribute is a good way to tell us this is on that image and we’ll get extra information from around your page with regard to how we can rank that landing page.”

Moz mentions ranking factors about alt text. Instead of saying that the alt text itself is a ranking factor, Moz advises:

“…alt text offers you another opportunity to include your target keyword. With on-page keyword usage still pulling weight as a search engine ranking factor, it’s in your best interest to create alt text that both describes the image and, if possible, includes a keyword or keyword phrase you’re targeting.”

In 2021, during a Twitter discussion about ALT text having a benefit on SEO, Google Developer Martin Splitt said:

“Yep, alt text is important for SEO too!”

Later in 2021, Mueller noted that alt text is not magic during a conversation about optimization for indexing purposes.

“My understanding was that alt attributes are required for HTML5 validation, so if you can’t use them with your platform, that sounds like a bug. That said, alt text isn’t a magic SEO bullet.”

[Recommended Read] → Ranking Factors: Systems, Signals, and Page Experience

Alt Text As A Ranking Factor: Our Verdict

Is Alt Text A Ranking Factor For Google Image Search?Is Alt Text A Ranking Factor For Google Image Search?

Alt text is a confirmed ranking factor for image search only. You should craft descriptive, non-spammy alt text to help your images appear in Google Image Search results.

Alt text is definitely not a ranking factor in Google Search. Google has clarified that alt text acts like normal page text in overall search. So it’s not useless, but it’s not a separately considered ranking factor in your page content.

That doesn’t mean you should ignore alt text. It’s a helpful accessibility tool for screen readers. When you’re writing alt text, ask yourself what you want someone who can’t see the image to understand about it.

Featured Image: Paulo Bobita/SearchEngineJournal

Source link

Keep an eye on what we are doing
Be the first to get latest updates and exclusive content straight to your email inbox.
We promise not to spam you. You can unsubscribe at any time.
Invalid email address
Continue Reading


Google On Traffic Metric & SEO




Google On Traffic Metric & SEO

Google’s recent Search Off the Record podcast discussed the important details of SEO, including a thought provoking segment about how SEOs can benefit from realign their thoughts about traffic with other possibly more important goals.

Traffic As A Measure Of Success?

One of the most misinformed things that search marketers do is measure their success by citing traffic statistics. We’ve all seen the posts on Facebook or the articles on blogs, where a search marketer relates how they did X for a client and traffic exploded exponentially within months.

Link builders use traffic as a metric of success, content writers do it, SEOs do it.

The question that goes through my mind when I see that is, well, what effect did that have on sales or ad clicks?

Because if earnings remain flat then the traffic increase probably doesn’t matter and probably neither did the SEO work done for that traffic.

Ask any Pay Per Click expert or any affiliate marketer about the importance of conversions versus traffic and the response is clear: conversions are everything, not traffic.

Adam J Humphreys, CEO of search marketing and design consultancy Making 8 (LinkedIn profile) said this about success metrics:

“I focused on analytics early on and learned the language of business which is ROAS. This is all executives care about other than awareness.

Our job is to grow their business first as well as awareness second.

Many don’t know how to measure attribution from SEO.

Things like form fills and calls are almost always unmeasured and unreported. Clients get super excited when I talk about this because it’s almost like nobody ever cared about their success.”

What Google Says About SEO & Traffic

Martin Splitt asked the question about the business impact of traffic.

Martin asked:

“…when you say traffic drops, what does that mean to you? Is it impressions?

Are you going to Search Console and click on the Performance report and you just look at the impressions going down or clicks?

Or do you actually measure a real business impact? Like you’re selling only half as many things as you used to sell last month?”

John Mueller added his thoughts on the topic. He tried to explain why SEOs don’t focus on return on investment (ROI) or the impact of SEO on earnings.

He offered his guess that maybe it’s the time between doing SEO versus the impact of it.

He commented:

“Usually what I see from people is that they focus on the traffic on their site.

And they look at something like Analytics and they say, ‘Oh, I get so many visitors and so many visitors from search engines. And that number went down significantly.’

…And the aspect of ROI or kind of the value of that traffic, I see that as something that a lot of these SEOs tend not to focus on primarily, because my assumption is it’s just a very long lead time there.

Like you can turn your website off now, but you might still have people who are kind of like paying for something for a while.

And then, it’s like, wow…”

Mueller followed up is thoughts on why SEOs might not focus on earnings by discussing how website traffic can be a misleading metric because it doesn’t tell a lot about why something happened, it only shows what happened.

He continued:

“But I do think it’s something where I sometimes feel it’s misleading to just purely focus on the traffic.

And I see that with our sites as well. A year or two ago, we would rank for the word “Google” in Canada or something like that, the search documentation, of course.

I hope the Google website ranks for the word Google. But like the Search documentation would rank for the word Google somewhere on the first page.

And we got tons of traffic there, but all of that traffic was basically irrelevant.

And then if you only look at the traffic and all of that irrelevant traffic goes away, then it might look like you lost a lot of traffic, but actually it’s like all of those things are people that weren’t relevant for your site anyway.

So you almost need to look at the bigger picture of all of the traffic that’s gone, but also keep in mind like, well, a lot of this was useless and I should maybe focus on the queries that people use, and then clicks and impressions for those, or individual like lower level pages of the site and kind of track those a little bit more.”

Traffic And SEO

SEOs promote themselves through case studies showing all the traffic their efforts brought. But those studies are hollow and maybe even deceptive if there’s absolutely no reference to how much sales lift resulted after their effort.

Link builders do a similar thing where they promote all the hundreds or thousands of links they acquired for a client and sometimes mention the lift in traffic. Just as consistently as SEOs, they always leave out the effect on sales or earnings.

Why do they do that? I suspect that many of those SEOs have never actually built a business around monetizing affiliate sales or ad clicks and thus don’t have first hand experience from that side of the SEO fence.

A lot of the old school SEOs like myself learned what we know from building and monetizing websites, maybe because client work for SEO wasn’t really as much of a thing then as it is now.

Using traffic as a metric is useful for measuring the impact of SEO but traffic should not be the goal of SEO. The reason is because there are different kinds of traffic.

Some traffic converts into sales. Some traffic has a lead time from the first visit to the sales. Some traffic is useful for building a brand name.

But some traffic is not relevant or useful.

When it comes to diagnosing traffic drops related to ranking changes, it may be helpful to understand if there’s any impact on sales and if not, to understand why the traffic drop had no monetary impact and if content and SEO efforts might be better directed in a different direction.

Listen to the Google podcast, Search Off the Record: SEO Is In The Details.

Source link

Keep an eye on what we are doing
Be the first to get latest updates and exclusive content straight to your email inbox.
We promise not to spam you. You can unsubscribe at any time.
Invalid email address
Continue Reading