CBPO

Monthly Archives: July 2019

How Google May Handle Question Answering when Facts are Missing

July 14, 2019 No Comments

I wrote about a similar patent in the post, Google Extracts Facts from the Web to Provide Fact Answers

This one introduces itself with the following statement, indicating a problem that Google may have with answering questions from the facts it may collect from the Web to fill its knowledge graph:

Embodiments relate to relational models of knowledge, such as a graph-based data store, can be used to provide answers to search queries. Such models describe real-world entities (people, places, things) as facts in the form of graph nodes and edges between the nodes. While such graphs may represent a significant amount of facts, even the largest graphs may be missing tens of millions of facts or may have incorrect facts. For example, relationships, edges or other attributes between two or more nodes can often be missing.

That is the problem that this new patent is intended to solve. The patent was filed in November of 2017. The earlier patent I linked to above was granted in June 2017. It does not anticipate missing or incorrect facts like this newer patent warns us about. The newer patent tells us about how they might be able to answer some questions without access to some facts.

It’s also reminding me of another patent that I recently wrote about on the Go Fish Digital Website. That post is titled, Question Answering Explaining Estimates of Missing Facts. Both the patent that post was about and this new patent include Gal Chechik, Yaniv Leviathan, Yoav Tzur, Eyal Segalis, as inventors (the other patent has a couple of additional inventors as well.)

The earlier question answering with estimates patent talks about how they might infer answers, and provide explanations with those answers. This also tells it might infer answers, but doesn’t include the explanations:

Facts and/or attributes missing from a relational model of knowledge often can be inferred based on other related facts (or elements of facts) in the graph. For example, a search system may learn that an individual’s grandfather is a male parent of a parent. Accordingly, the system can determine with high confidence that an individual’s grandfather, even though there is no grandfather edge between nodes, is most likely a parent of a parent (given that there is a parent edge between nodes) with an additional check the parent of the parent is male. While this example uses one piece of supporting evidence (called a feature), inferring an individual’s grandfather, functions estimating missing facts are often more complex and can be based on several, even hundreds, of such features. Once the facts and/or attributes missing from a relational model of knowledge can be inferred, queries based on the facts and/or attributes missing from a relational model of knowledge can be resolved.

The process described in this question answering patent describes how Google may go about coming up with an answer to a question. This patent was filed after the one that includes estimates of how answers were created, so it does not include that step:

In one example embodiment, a computer system includes at least one processor and a memory storing a data graph and instructions. The instructions, when executed by the at least one processor, cause the system to generate a template sentence based on a fact including a first node, a second node and a string, wherein the first node and the second node exist in the data graph and the string represents a fact that is absent from the data graph, search the internet for a document including the template sentence, and upon determining the internet includes the document with the template sentence, infer the fact by generating a series of connections between nodes and edges of the data graph that together with the first node and the second node are configured to represent the fact, the series of connections defining a path, in the data graph, from the first node to the second node.

This process isn’t described in too much detail, but the patent does provide an example, which may be helpful in understanding how it may work. Here is that example:

For example, a node may correspond to a fact describing a parent-child relationship. For example, baseball player Bob Boone is the son of baseball player Ray Boone and the father of baseball players Aaron Boone and Bret Boone. Accordingly, the data graph may include an entity as a node corresponding to Bob Boone, which may include an edge for a parent relationship directed to Ray Boone and two edges for child corresponding, respectively, to Aaron Boone and Bret Boone. The entity or node may also be associated with a fact or an attribute that includes an edge (e.g., occupation) between Bob Boone as a node and baseball as a node. Alternatively, the node Bob Boone may include an attribute as a property (e.g., occupation) set to baseball.

However, there may be no edge in the entity (or the graph as a whole) corresponding to a grandparent relationship. Therefore, the relationship between Ray Boone and Aaron Boone may not be shown in the graph. However, the relationship between Ray Boone and Aaron Boone may be inferred from the graph so long as the question answering system knows (i.e., has been instructed accordingly) that there is such an entity as a grandparent.

The inference may be based on the joint distribution of one or more features, which represent facts in the data graph that are related to the missing information. The system may also be used to store the inferences (e.g., as functions or algorithms) and the semantically structured sentence (e.g., X is the attribute of Y) used to generate the inference. It then uses these entities to map new string that corresponds to relationships between nodes. By that system may be configured to learn new edges between existing nodes in the data graph. In some implementations, the system can generate an inference and its algorithm from a very large data graph, e.g., one with millions of entities and even more edges. The algorithm (or function) can include a series of connections between nodes and edges of the data graph. Accordingly, the algorithm can represent an attribute as an edge in a fact. The algorithm (or function) can also include a check of a property of a node (e.g., a gender property is male). While the system in FIG. 1 is described as an Internet search system, other configurations and applications may be used. For example, the system may be used in any circumstance where estimates based on features of a joint distribution are generated.

The mentions of Joint Distributions in this patent are worth studying in more depth as the relationships between properties of different entities may reveal information that worth a system like the knowledge graph knowing about. The son of someone’s son is their grandson. If the knowledge graph doesn’t include that grandson property, then being able to make that connection can mean that a question answering system can start answering questions like Aaron Boone is Ray Boone’s Grandson. Other relations beyond whom is related to whom within a family can use this approach to answer questions as well.

This patent that is aimed at helping fill in missing and incorrect facts for question answering systems is:

Semi structured question answering system
Inventors: Yaniv Leviathan, Eyal Segalis, Yoav Tzur, and Gal Chechik
Assignee: GOOGLE LLC
US Patent: 10,346,485
Granted: July 9, 2019
Filed: November 8, 2017

Abstract

In one example embodiment, a computer system includes at least one processor and a memory storing a data graph and instructions. The instructions, when executed by the at least one processor, cause the system to generate a template sentence based on a fact including a first node, a second node and a string, wherein the first node and the second node exist in the data graph and the string represents a fact that is absent from the data graph, search the internet for a document including the template sentence, and upon determining the internet includes the document with the template sentence, infer the fact by generating a series of connections between nodes and edges of the data graph that together with the first node and the second node are configured to represent the fact, the series of connections defining a path, in the data graph, from the first node to the second node.

Some posts I’ve written about patents involving question answering:

L:ast Update July 11, 2019.


Copyright © 2019 SEO by the Sea ⚓. This Feed is for personal non-commercial use only. If you are not reading this material in your news aggregator, the site you are looking at may be guilty of copyright infringement. Please contact SEO by the Sea, so we can take appropriate action immediately.
Plugin by Taragana

The post How Google May Handle Question Answering when Facts are Missing appeared first on SEO by the Sea ⚓.


SEO by the Sea ⚓


Facebook reportedly gets a $5 billion slap on the wrist from the FTC

July 14, 2019 No Comments

The U.S. Federal Trade Commission has reportedly agreed to end its latest probe into Facebook‘s privacy problems with a $ 5 billion payout.

According to The Wall Street Journal, the 3-2, party-line vote by FTC commissioners was carried by the Republican majority and will be moved to the Justice Department’s civil division to be finalized.

A $ 5 billion payout seems like a significant sum, but Facebook had already set aside $ 3 billion to cover the cost of the settlement and the company could likely make up the figure in less than a quarter of revenue (the company’s revenue for the last fiscal quarter was roughly $ 15 billion). Indeed, Facebook said in April that it expected to pay up to $ 5 billion to end the government’s probe.

The settlement will also include government restrictions on how Facebook treats user privacy, according to the Journal.

We have reached out to the FTC and Facebook for comment and will update this story when we hear back.

Ultimately, the partisan divide which held up the settlement broke down with Republican members of the commission overriding Democratic concerns for greater oversight of the social media giant.

Lawmakers have been calling consistently for greater regulatory oversight of Facebook — and even a legislative push to break up the company — since the revelation of the company’s mishandling of the private data of millions of Facebook users during the run up to the 2016 presidential election, which wound up being collected improperly by Cambridge Analytica.

Specifically the FTC was examining whether the data breach violated a 2012 consent decree which saw Facebook committing to engage in better privacy protection of user data.

Facebook’s woes didn’t end with Cambridge Analytica . The company has since been on the receiving end of a number of exposes around the use and abuse of its customers’ information and comes as calls to break up the big tech companies have only grown louder.

The settlement could also be a way for the company to buy its way out of more strict oversight as it faces investigations into its potentially anti-competitive business practices and inquiries into its launch of a new cryptocurrency — Libra — which is being touted as an electronic currency for Facebook users largely divorced from governmental monetary policy.

Potential sanctions proposed by lawmakers for the FTC were reported to include the possibility of elevating privacy oversight to the company’s board of directors and potentially the deletion of tracking data; restricting certain information collection; limiting ad targeting; and restricting the flow of user data among different Facebook business units.


Social – TechCrunch


Space Photos of the Week: A Tribute to Voyager’s Twin Trippers

July 14, 2019 No Comments

These two missions fundamentally changed our understanding of the solar system. See how in this entrancing photo gallery.
Feed: All Latest


Why Honesty is the Best Policy in PPC Marketing

July 13, 2019 No Comments

Over the years, I have seen so many horror stories when it comes to PPC Management. Whether it’s advertisers flying blind with their ad budgets or the common event of not knowing that their ads are being shown with irrelevant terms, there should always 100% transparency between the agency and the client. Furthermore, there needs to be more HONESTY on behalf of the PPC Agency. In this post, I will talk about a few areas of the Agency/Client Relationship that should be based on being honest with the client.

Educate the Advertiser:

Let’s face it, the PPC agency knows more about PPC Marketing than the client. However, that does not mean the client needs to be taken advantage of because they do not know how everything works. The person handling the client’s account needs to “in many ways” educate the client as to what is working, not working and where there are opportunities.

Admit Mistakes:

Everyone makes mistakes, right? Well, PPC Agencies should not try and hide them just because they can get away with it. Agencies should be forthcoming with admitting mistakes that were made and how efficiently and effectively they were fixed. It’s better to be honest with the client, than having them find out later that you lied to them. Ever heard of a Referral or a Testimonial?

Honest and Factual Reporting:

Over the years, I have seen so many poor examples of PPC Reporting where clients receive an excel spreadsheet of just Clicks, Impressions, CTR%, CPCs, etc… and not a single keyword or text ad or even a sentence on the performance of the account. In today’s world that is unacceptable.  Moreover, I have also seen examples of trend charts being manipulated to disguise the true performance of a specific metric. Agencies have a responsibility to provide not only excellent service, but also honest and factual reporting.

Managing Expectations:

PPC Marketing is not for everyone and for those who are spending money have this perception that the more they spend the better the results. That is completely FALSE. If an client/advertiser was given any sort of Guarantee from an agency, they should “run for the hills”. Guarantees in PPC Marketing are very dangerous for both parties because they create false expectations. An agency must be honest and upfront with the client when it comes to setting expectations both on performance and future success. The agency must have a clear understanding of the client’s:

  • Cost per Conversions/Acquisition
  • Targeted Audience
  • Messaging Tactics
  • Daily and Monthly Budgets

In Conclusion:

Honesty is always the best policy in PPC. Agencies have a responsibility to not only provide excellent service, but also be honest and forthcoming with the client. I have heard countless stories of poor PPC Management, including the topics I mentioned in this post. Some may say that is good for the industry because it creates more “turnover” and more opportunities for other agencies. However, for this PPC Geek, I believe in Happy Clients.


PPC Marketing Consultant | Google Ads Agency


Judge dismisses Oracle lawsuit over $10B Pentagon JEDI cloud contract

July 13, 2019 No Comments

Oracle has been complaining about the procurement process around the Pentagon’s $ 10 billion, decade-long JEDI cloud contract, even before the DoD opened requests for proposals last year. It went so far as to file a lawsuit in December, claiming a potential conflict of interest on the part of a procurement team member. Today, that case was dismissed in federal court.

In dismissing the case, Federal Claims Court Senior Judge Eric Bruggink ruled that the company had failed to prove a conflict in the procurement process, something the DOD’s own internal audits found in two separate investigations. Judge Bruggink ultimately agreed with the DoD’s findings:

We conclude as well that the contracting officer’s findings that an organizational conflict of interest does not exist and that individual conflicts of interest did not impact the procurement, were not arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law. Plaintiff’s motion for judgment on the administrative record is therefore denied.

The company previously had filed a failed protest with the Government Accountability Office (GAO), which also ruled that the procurement process was fair and didn’t favor any particular vendor. Oracle had claimed that the process was designed to favor cloud market leader AWS.

It’s worth noting that the employee in question was a former AWS employee. AWS joined the lawsuit as part of the legal process, stating at the time in the legal motion, “Oracle’s Complaint specifically alleges conflicts of interest involving AWS. Thus, AWS has direct and substantial economic interests at stake in this case, and its disposition clearly could impair those interests.”

Today’s ruling opens the door for the announcement of a winner of the $ 10 billion contract, as early as next month. The DoD previously announced that it had chosen Microsoft and Amazon as the two finalists for the winner-take-all bid.


Enterprise – TechCrunch


How to grab featured snippet rankings with zero link building effort

July 13, 2019 No Comments

Featured snippets, also known as “position zero” placements on Google, have been receiving their fair share of glory and blame lately. 

While some big corporations like Forbes went ahead and questioned if Google is stealing traffic with the featured snippet, content creators like me have found it easy to get more traffic, thanks to being able to rank small sites on a featured snippet.

This post will give you a brief idea on how you can rank a page on Google’s featured snippet — without building any links to that page.

Understand the types

There are three major types of featured snippets that you can go for. As most of our clients are bloggers, we tend to go for either the paragraph snippets or the list snippets. Table snippet is another popular one that you can target.

Here’s a quick graph from Ahrefs about the snippet type and their percentages.

graph about the snippet type and their percentages

Targeting the right keywords

Once you finalize the type of snippet that you would want to go for, it is time to dig deep into your keyword research to find keywords that suit your blog and match the requirements for the type of snippet that you are going after.

If you are going for a paragraph snippet, you will have to find keywords that are primarily related to these types:

  • How to
  • Who/what/why

example of finding keywords on snippets

If you are trying to rank for a numeric list (numbered list or bullet points), the idea would be to structure your content in a way so that it offers step by step guides to someone. As per our experience, Google only shows a numeric list on featured snippet when the keyword tells Google that the searcher is looking for a list.

example of a listed featured snippet

For table snippets, the idea is to have structured schema data on your website that compares at least two sets of data on the page. You don’t really have to have a properly formatted column-based table to be able to rank for table snippets as long as the comparison and the schema is there.

example of a table structured snippet

Understanding the type and targeting the right keywords will do more than half of the job for you when it comes to ranking your website on the featured snippet with zero links.

However, you are not going to win the battle by out-throwing an already existing featured snippet. This will only work for keywords that don’t already have a featured snippet ranking on Google.

To grab featured snippets from the existing competition, you will need to go ahead and perform a few more steps.

Copying your competitor

Some will call it “being inspired”, but essentially, what you are doing is copying the structure of an existing featured snippet article and trying to make it better (both with content and if possible, with links).

What do I mean when I say, copying the structure of an existing page and making it better? If you want to rank for the featured snippet for the keyword “best cat food brands” and if the one, ranking at this moment already has a list of 20, you will have to create a list of 25, in the exact same format that the current one is using.

Once that’s done, the final step is simply to make sure you have proper schema on the page.

Note: It is very unlikely that this method will help you outrank an existing featured snippet unless you also rank in the top ten for that keyword.

How do we find keywords for featured snippets?

As you can imagine, finding the right keyword to target is winning half of the battle when it comes to ranking on featured snippets.

I use Semrush, but feel free to use your own tools. Here’s what our agency’s process looks like.

Let’s assume, for the purpose of this article, that I run a pet blog and I am interested in ranking for multiple featured snippets.

I would go to Semrush, and put one of my competitors on search.

example of competitor research on semrush

Source: semrush

Now click on “Organic Research”, select positions and from advanced filters, select – Include > Search features > featured snippet.

example of organic research

Source: semrush

This will give you a huge list of keywords that are currently ranking as featured snippets. As you can see, we found about 231 opportunities to target here:

listing of potential keywords for targeting

Source: semrush

It is time to add another condition to our advanced filters. Let’s select include > words count > greater than five. Here’s what the new result looks like:

example of using advanced filters in semrush

Source: SEMrush

From here on, simply organize the keywords by volume and then select the ones that you think matches with your target market. Like any keyword research, you will have to find keywords that have low competition and moderate search volume. Personally, I would try to go for keywords that have less than 500 monthly searches.

Make sure that you are following the initial three steps that we discussed. You will almost always have a higher chance of ranking on featured snippet following this strategy.

Khalid Farhan blogs about internet marketing at KhalidFarhan.com. He can be found on Twitter @iamkhalidfarhan.

The post How to grab featured snippet rankings with zero link building effort appeared first on Search Engine Watch.

Search Engine Watch


Twitter will start testing its ‘hide replies’ feature next week, in Canada

July 12, 2019 No Comments

Twitter users are getting more control over which comments are visible in the conversations they start.

The company has been testing and talking about this feature since earlier this year, but starting next week, Twitter will actually roll it out to users in Canada.

As you can see in the GIF below, when you’re looking at replies to your tweets, you’ll be able to select any of them and hit the “hide reply” option. However, as the name implies, these posts won’t be fully removed from Twitter, just hidden from the default view — everyone will still be able to tap on a gray icon to view hidden replies.

Here’s how Twitter’s Michelle Yasmeen Haq and Brittany Forks explain the feature:

Everyday, people start important conversations on Twitter, from #MeToo and #BlackLivesMatter, to discussions around #NBAFinals or their favorite television shows. These conversations bring people together to debate, learn, and laugh. That said we know that distracting, irrelevant, and offensive replies can derail the discussions that people want to have. We believe people should have some control over the conversations they start.

Twitter Hide Replies

As my colleague Sarah Perez noted previously, the current implementation is open to at least two criticisms — one, that it could allow users to hide critical viewpoints or fact-checking of their tweets (maybe quote-tweeting will be the better strategy moving forward), and two, that it still forces people to wade through potentially trollish or hateful content in order to hide replies.

Haq and Forks emphasize that Twitter is still looking for ways to improve the feature: “By testing in one country we want to get feedback and better understand how this tool can improve before it’s available globally.”

And yes, the timing of the news is a little awkward, coming right after Twitter went down for about an hour.


Social – TechCrunch


Why Dogs Now Play a Big Role in Human Cancer Research

July 12, 2019 No Comments

There’s a strong chance your aging dog will get cancer—but your pupper could also help humans survive it.
Feed: All Latest


#Ham4PPC: A Tribute to Digital Marketing

July 12, 2019 No Comments

Like anything, mastering the fundamentals of pay-per-click account management requires both practice and perseverance. Today’s #Ham4PPC comprises nine lessons I’ve learned from my years as a PPC professional.

Read more at PPCHero.com
PPC Hero


Snapchat announces new shows from Serena Williams, Arnold Schwarzenegger and others

July 10, 2019 No Comments

Snapchat just announced that it’s making shows with big names like Serena Williams, Arnold Schwarzenegger and Kevin Hart, as well as online stars like Emma Chamberlain, Loren Gray, Rickey Thompson, Baby Ariel and FaZe Banks.

Snapchat launched its original content efforts two years ago, and today it’s unveiling a new program called Creator Shows. As initially announced in the Hollywood Reporter, these will be first-person shows designed around individual creators.

For example, Schwarzenegger will be providing motivational advice in a show called “Rules of Success,” while Thompson will weigh in on fashion and lifestyle trends on “Trend or End” and Gray offers beauty advice on “Glow Up.”

The shows will begin airing this month. They’re all exclusive to Snapchat, and many of them come from creators who have a substantial following on other platforms — Chamberlain, for example, was just described in The New York Times as “the funniest person on YouTube.

Rickey Thompson Premieres July 10

“Snapchat has always been my favorite platform to post random and funny things on because it’s so relaxed,” Chamberlain said in a statement. “My favorite part about it is that I get to watch my own Snapchat Stories a few hours after I post them for entertainment… kind of embarrassing, I know…”

Snap isn’t sharing viewership numbers around its original shows, but it does say that daily time spent watching those shows tripled over the past year.

And as media giants funnel more and more money into original video content, this might be the strategy that Snapchat needs to compete — rather than trying to find the next big-budget hit, it can focus on personality-driven shows from creators with large followings.


Social – TechCrunch