Leo learns from the community

Leo
We’re designing systems to protect against machine learning bias

In the wake of recent acts of extreme brutality and injustice and mass protests, we’re examining our role in perpetuating systems of inequality. We are responsible for our impact as a tech company, as a news reader, and, acutely, as a developer of machine learning algorithms for Leo, your AI research assistant. 

Artificial intelligence and machine learning are powerful tools that allow Leo to read thousands of articles published every day and prioritize a top selection based on the topics, organizations, and trends that matter to you. However, if not designed intentionally, these tools run the risk of reinforcing harmful cultural biases.

Bias sneaks into machine learning algorithms by way of incomplete or imbalanced training data. Without realizing it, we miss or overrepresent certain variables and the algorithm learns the wrong information, often with dangerous outcomes.

In the case of Leo, we risk introducing bias when teaching him broad topics such as “leadership.” Leo learns these topics by finding common themes in sets of articles curated by the Feedly team. For the topic “leadership,” Leo might pick out themes like strong management skills and building a supportive team culture. However, if more articles about male leaders than female are published or added to the training set, Leo might also learn that being male is a quality of leadership. Tracking which themes Leo learns is an essential part of topic modeling that helps prevent us from reinforcing our biases or those of the article author or publisher.

It’s on us as developers to be deliberate and transparent about the way we account for bias in our training process. With that in mind, we’re excited to share what we’re working on to reduce bias at the most crucial stage: the training data

Break down silos

Collaboration among folks from diverse backgrounds helps us account for our blind spots. However, to make that collaboration possible, we need an accessible tool. The new topic modeler is that tool — designed so that anyone in the Feedly community can help curate a dataset to train Leo about topics they’re passionate about.

A peek inside the topic modeler tool

The topic modeler takes advantage of the Feedly UI we know and love to allow multiple users to search for articles for the training set and review Leo’s learning progress. Our goal is to connect with experts in a variety of fields to build robust topics that represent our entire community — not just the engineering team.

Put to the test: the diversity topic

Recently, two Feedly team members with no machine learning experience and who are interested in diversity issues road tested the new tool to redesign our diversity topic. The result is a topic that is rich and nuanced: rather than focusing only on the buzzword “diversity,” Leo will be looking for thousands of related keywords, including representation, inclusion, bias, discrimination, equal rights, and intersectionality. Now you can train Leo to track diversity and inclusion progress in your industry and find essential information for how to build and maintain inclusive work cultures and hiring practices.

Leo prioritizes diversity in your Science feed

Leo continuously learns

Topic modeling is not the only way to collaborate. Any Feedly user can help Leo learn. When Leo is wrong, you can use the ‘Less Like This’ down arrow button to let him know that an article he’s prioritized isn’t about a particular subject.

Leo will also seek your feedback occasionally via a prompt at the top of an article. If you see “Is this article about [topic]?,” let him know! Your feedback gets incorporated into Leo’s training set to fill in any gaps we missed and strengthen his understanding.

Your feedback helps fine-tune Leo’s understanding

Join the movement

Beyond in-app feedback, feel free to reach out via email or join the Feedly Community Slack channel, especially if you have a topic for Leo to learn about. This is the tip of the iceberg when it comes to addressing and dismantling systemic bias. We take our role as content mediators seriously and know that we are indebted to those who have fought for so long to bring these issues to our attention. Leo is listening and learning.

Leo recognizes pharmaceutical drugs from psychoactive drugs

We heard from lots of biopharma users that the Drugs topic could be improved and clarified, considering the different meanings it has.

We are excited to announce that you can now prioritize either pharmaceutical drugs or recreational drugs.

We have taught Leo to understand if an article is about pharma drugs or psychoactive ones to improve the relevance of his prioritization.

Let me show you how it works.

Pharmaceutical Drugs

Let’s imagine that you have a Biopharma Business feed and want to track updates about drugs treating cancer. Let’s train Leo to read this feed and cut through the noise for you.

Click ‘Train Leo’ and search for the new #Drugs (pharmaceutical) topic
Leo knows how to recognize articles about pharmaceutical drugs

You can see a preview of all the articles that Leo has recognized as related to pharmaceutical drugs and cancer. These articles will be prioritized in your feed.

Psychoactive Drugs

Now, you can do the same with psychoactive drugs. Let’s say you want to prioritize articles that are at the intersection of psychoactive drugs and mental health.

Search for the new #Drugs (psychoactive) Leo topic

Leo will continuously learn and get smarter as he prioritizes articles about pharma drugs or recreational drugs, letting you focus on the topics and trends that matter to you.

After two months of Leo utilization, I can say that he saved us two of the three hours that we needed weekly to do our job, with the same or better quality. Really performant. Good job and long life to Leo 😉

Jessyca Duer, UnitedHealth Group

Train your Leo now

We are excited to see many Feedly users declutter their feeds and dig deeper into the topics and trends that matter to them. Sign up today and discover what Leo can do for you!

If you are interested in learning more about Leo’s roadmap, you can join the Feedly Community Slack. 2020 will be a challenging year, but by staying informed, you can respond better and remain in control. 

Leo understands malware threats

Research and prepare for the latest malware threats without the information overload

Cybersecurity is a game of foresight. It’s a chessboard on which attackers and defenders are constantly looking for checkmate. 

Hackers launch a new ransomware attack every 14 seconds. They’re increasingly more capable and sophisticated. Learning how they plan attacks, what techniques they use, and who they’re targeting, can make you so much better prepared. You’ll save the cost and headache of a cyber assault too. This is especially important considering that the cost of ransomware attacks in the U.S. alone surpassed $7.5 billion in 2019.

But investigating malware threats is tedious. Hundreds of new articles and tweets need to be reviewed and triaged every day. Finding critical threats in that sea of information is time-consuming and overwhelming.

We want to help you streamline your tactical and operational open-source intelligence, so that you can better protect your environment.

That’s why we’ve taught Leo, your AI research assistant, to recognize malware threats. You can ask him to read your security feeds and prioritize what’s relevant to you, your sector, and your environment.

Let’s imagine that you work in a threat intelligence team and are responsible for researching and analyzing the threat landscape. You’re particularly interested in evolving malware threats (including ransomware and malvertisement).

Cut through the noise

You can train Leo to read your Security News feed and prioritize articles related to malware.

Leo prioritizes malware articles in your Security News feed

Leo continuously reads the thousands of articles published in those feeds. It’s an efficient way to cut through the noise and keep up with the evolving malware landscape without the overwhelm.

You’re in control

Leo has been trained to understand broad topics like malware, as well as hundreds of specific malware types like malvertisement, ransomware, adware, bots, rootkits, spyware, etc.

Asking Leo to prioritize malware in your Security News feed is as simple as creating a new Topic priority and selecting ‘malware’ as the topic.

Ask Leo to prioritize malware threats in your Security News feed

You can combine topics with +AND and +OR and create even more targeted priorities for Leo. For example, use +AND to focus on malware related to Android or top companies in your sector.

Refine the priority to malware and Android

You can also ask Leo to look for a specific type of malware like malvertisement or ransomware.

Prioritize ransomware threats

Continuously learning and getting smarter

Leo is smart. He continuously learns from your feedback. When Leo is wrong, you can use the ‘Less Like This’ down arrow button to let him know that an article he’s prioritized isn’t about malware.

Let Leo know when he’s wrong

Break down silos

Bring your research team into the picture. They can create a Threat Intel Report Board and save the most critical insights they discover in their Feedly. Then everyone with the same Board can leave notes and highlight the biggest threats. 

We’ve seen teams create tactical and operational Boards. For instance, a Vulnerability Report can be built up with information for those that deal with security procedures, while strategic CISO Newsletters can keep management up to speed about malware and your planned response.

Articles bookmarked in a Board can be shared with the rest of the team via daily newsletters, Slack and Microsoft Teams notifications, or pushed to other apps using the Feedly Cybersecurity API.

Share the threat intelligence you collect in Feedly with other teams and apps

Streamline your open-source intelligence

We’re excited to see how your security team will declutter your feeds and dig deeper into the critical threats that matter to you. Sign up today and discover Feedly for Cybersecurity.

If you’re interested in learning more about Leo’s roadmap, you can join the Feedly Community Slack channel. 2020 will be a thrilling year with new skills and bold experiments!

Leo understands threat actor groups

Research threat actor groups and learn more about their tactics, techniques, and procedures without the overwhelm

Cyber attacks continue to wreak havoc around the world. The actors waging these wars don’t just care about fraud either. They’re part of criminal organisations. Foreign governments stealing data for defense or national interests. Even terrorists or activists driven to disrupt and cause harm. 

What’s more, they’re increasingly capable and sophisticated. It’s a growing threat that can strike anyone at any time.

When you learn about threat actors’ tactics and motivations, you can better prepare against them, saving you the costs and headaches that come with a breach or attack. 

But there’s so much content to wade through when investigating these threat actors. It’s like fishing blind in an ocean. You’ll never know what’s coming back on the hook. More time and stress is spent on finding information about the threat, rather than acting on it. You can be overwhelmed. 

We’re passionate about helping you refine and streamline your open-source intelligence. That’s why we’ve taught Leo, your AI research assistant, to recognize threat actor groups. He can find them in your Feedly security feeds, prioritizing articles related to the actors and sectors you care about.

Let’s imagine that you work in the telecommunications sector, and you’re researching the tactics and motivations of MuddyWater, an Iranian threat actor group.

Cut through the noise

You can train Leo to read all your cybersecurity, foreign affairs, and cyber warfare sources, and prioritize articles related to MuddyWater.

Prioritize a threat actor

Leo continuously reads the articles in your feeds and prioritizes the ones that mention MuddyWater (or any of its aliases). It’s a powerful and effective way to keep up with their latest techniques, tactics, and procedures.

You’re in control

Leo has been trained to recognize all the threat actor groups referenced by the MITRE ATT&CK framework. This is a list of common names for hacking groups, as recognized by the global security community.

Asking Leo to prioritize MuddyWater in your security feed is as simple as creating a new Topic priority and selecting ‘MuddyWater’ as the topic.

Enter a threat actor alias in the topic field

When you prioritize MuddyWater, Leo will also look for other synonyms for that group like Seedworm and TEMP.Zagros.

You can combine topics with +AND and +OR to create even more targeted priorities for Leo. For example, use +AND to combine an actor group with an attack vector or a sector. This narrows his focus further so you find exactly what you’re looking for.

Continuously learning and getting smarter

Because Leo is integrated with the MITRE ATT&CK framework, it’s continuously learning and getting smarter. As new groups or aliases are identified, they’ll be automatically updated in your Feedly.

Leo recognizes threat actor groups listed on the MITRE ATT&CK framework

Break down silos

As you search and discover new content, share insights with your research team. Together, you can create a Threat Intel Report Feedly Board and bookmark the most critical insights you discover. You can also add notes and highlights about why a threat is high-priority.

We’ve already seen security teams create tactical Boards, such as a Vulnerability Report, to share with their operations experts. You might also want to build a CISO Newsletter to keep your management updated. It’s all possible within Feedly.  

Articles bookmarked in a Board can be shared with the rest of the team via daily newsletters, Slack or Microsoft Teams notifications, or pushed to other apps using the Feedly Cybersecurity API.

Share the threat intelligence you collect in Feedly with other teams and apps

Streamline your open-source intelligence

We’re excited to see how your security team will declutter your feeds and dig deeper into the critical threats that matter to you. Sign up today and discover Feedly for Cybersecurity.

If you’re interested in learning more about Leo’s roadmap, you can join the Feedly Community Slack channel. 2020 will be a thrilling year with new skills and bold experiments!

The Feedly Cybersecurity API

Feedly for Cybersecurity includes an API that allows cybersecurity teams to share the intelligence they collect in Feedly with other applications

150,000 cybersecurity professionals use Feedly to collect intelligence about the evolving threat landscape. 

Threat research and collection are one step of the overall threat intelligence, investigation, and response.

The Feedly Cybersecurity API allows security teams to easily integrate the insights they collect in Feedly into other systems and applications. Some teams use the API to extract data about threats and vulnerabilities and feed larger machine learning threat-prioritization models. Some teams use the API to create Jira tickets based on the content of the Feedly boards to make sure that critical vulnerabilities are reviews and patched in a timely manner.

Access to the Feedly API (up to 200,000 requests per month) is an add-on included in the Enterprise Edition of the Feedly for Cybersecurity package.

In this tutorial, we will show you how to use the Feedly API to access the content of your security feeds, your boards, and your Leo priorities.

Authentication

When you subscribe to Feedly for Cybersecurity Enterprise Edition, we will provide you with a special Feedly access token associated with your account. That token will allow you to access the content of your feeds, boards, and priorities and perform up to 200,000 requests per month.

Articles as JSON

The JSON representation of an article combines some of the open-source content included on the RSS or on the website, CVE/CVSS/Exploit information aggregated from vulnerability and exploit databases, as well as the results of the Leo cybersecurity models.

The title, content, and visual information give you access to the core of the content of the articles:

JSON representation of the core of the article

The commonTopics array represents Leo’s topic classification. The entities represent CVEs, products, or companies Leo has identified in the article. The CVE entity includes CVSS and exploits information extracted from vulnerability databases.

The estimatedCVSS represents the result of Leo’s CVSS scoring model. This is useful for zero-days and articles which do not mention a CVE explicitly. In those cases, Leo reads the content of the article and computes an approximative CVSS score based on the terminology used in the article or the tweet.

Leo enrichment of the article

Pro tip: When you have an article open in the Feedly web application, you can use the Shift+D keyboard shortcut to see and inspect the JSON of the article.

Use keyboard shortcut SHIFT+D to see the preview of the article JSON

Accessing the content of your feeds

Let’s imagine that you have a “Security News” feed which contains a list of known and trusted security sources you want to follow.

The Feedly API allows you to query Feedly and ask for the last 100 articles aggregated in that feed. The articles are normalized in a JSON format which includes the title, the content, the source information, as well as all some cybersecurity metadata (Leo topics classification, CVE metadata, CVSS metadata, exploit information.

You can use the Stream endpoint to get the last 100 articles published in a feed:

Overview of the stream endpoint

The most important parameter is the streamId. Each feed in your Feedly account has a unique stream id. When you select the feed in the left navigation bar, you see the streamId as part of the URL. The stream id is formatted as `enterprise/xxxx/category/xxxx` for team feeds and `user/xxxx/category/xxxx` for personal feeds.

Finding the streamId of a feed

The count parameter defines the number of articles the server will return. We recommend that you select a number between 20 and 100. If you need access to more than 100 articles, you can use the continuation parameter returned by the response to chain the requests and ask for the next 100 articles.

Finally, the importantOnly parameter allows you to get the list of articles in the stream that has been prioritized by Leo.

Troubleshooting tips:

  • Make sure that the requests you are making are authenticated using the token you have received from the Feedly team.
  • Make sure that the streamId is URL encoded when it is passed as a parameter to the Stream endpoint.

Accessing the content of your boards

Security teams use boards to bookmark critical articles everyone in the team should be aware of. They also often use boards to bookmark articles they want to share with other applications.

You can use the same Stream endpoint to access the last N articles manually bookmarked by your team to a board.

The only difference will be the streamId. Team Board streamIds are formatted as `enterprise/xxxx/tag/xxxx`. Personal Board streamIds are formatted as `user/xxxx/tag/xxxx`.

Finding the streamId of a board

If users have annotated the articles with some notes and highlights while saving the article to a board, those notes and highlights will be included in the article JSON structure.

JSON of notes and highlights

Example: Integrating Feedly with your ticketing system

Here is an example of how you can streamline the integration between the research and collection work of your threat intelligence team and the analysis and patching work of your operations team.

The research team creates a Feedly board called Critical Vulns where why bookmark articles related to critical vulnerabilities they want the operations team to be aware off and review.

Each time the research team finds a critical insight, they save that article in the Critical Vulns board, adding a note about why they think the vulnerability needs to be reviewed and patched.

Instead of asking the research team to manually create a ticket in your ticketing system (Jira, Service Now, etc.), you can write a small app which every 5 minutes connect to the Critical Vulns board, requests the last 20 articles bookmarked in that board, and for each new article, used the API of your ticketing system to create a new ticket. The app can enrich the ticket with the URL of the article saved in the board, the CVE information, and the notes and highlights from the researcher.

This is a powerful way to break the silos between your research team and your operations team and make sure that critical vulnerabilities are patched faster.

Pro tip: there is a simple solution to finding the new articles saved in a board. When your app processes a list of articles, it should save the first article in the list and the next time it uses the Stream Feedly app to get the latest articles bookmarked to a board, your app can use the newerThan parameter of the /v3/stream/content and pass that article id instead of a timestamp to get newer articles.

A lot more…

The Feedly web application and mobile applications are built on top of the Feedly API. This means that every piece of information available in the application and every action taken in the application is available in the API.

For more information about the Feedly API, please visit the Feedly Developer Website.

Streamline your open-source intelligence

We are excited to see many security teams use the Feedly API to streamline their open-source threat intelligence process. Sign up today and discover what Feedly for Cybersecurity can do for you!

If you are interested in learning more about Leo’s roadmap, you can join the Feedly Community Slack. 2020 will be a thrilling year with new skills and bold experiments!

Leo Understands COVID-19

Look beyond the big headlines. Leo can show you exactly what’s happening to your industry as a result of COVID-19, or filter it out.

Coronavirus news is everywhere right now. It’s not so much a wave of information as an ocean. It’s easy to get overwhelmed or miss a crucial market development. 

Or maybe you want to cut out the COVID-19 content altogether so you can find out what else is happening around the world. 

So we’ve taught Leo, your AI research assistant, how to help.

Mute or prioritize COVID-19 in your Feedly

Leo can already learn what you like to see and refine your Feedly. Now, he can mute or prioritize COVID-19 as well. And he does it across tens of millions of trusted sources. 

It works just like Leo’s other prioritization parameters such as keywords, topics, and events. ‘Coronavirus’ and ‘COVID-19’ are just two of the terms he recognizes. Leo takes into account a variety of the virus’s other names, too, like SARS-CoV-2. 

Leo prioritizes mentions of COVID-19 and its wide variety of aliases

Once you give Leo a priority, you’ll get a specific view of how your industry is reacting to the pandemic. Then just save the most interesting publications in your Feedly Board. 

You can mute or prioritize one feed, or every feed, and those feeds can be personal or spread across your team. It lets some team members focus on COVID-19 news if they need to, while others look beyond it. 

Here’s a few examples to show how Leo’s coronavirus filter might work for you. After all, the virus is impacting every sector, whether you’re in retail, cyberspace, automotive or pharmaceuticals…

COVID-19 and biopharma

You’re a drug development director looking for news and insight around cardiovascular disease, and how COVID-19 is affecting this research. 

Let’s imagine you have a Cardiology feed in Feedly, and you’re following multiple science and medicine journals. Hit ‘Train Leo’ in the top left toolbar. You can prioritize COVID-19 subjects by entering it as a topic.

Preview the prioritized COVID-19 articles in your Cardiology feed

The publications displayed are now all about coronavirus and cardiology. 

Refine the search further with +AND or +OR. Here’s some more information about Leo’s topic combinations.

COVID-19 and cybersecurity

You’re part of a large tech company. Security threats may have emerged during the pandemic, buried by the noise online. 

Do the exact same thing. Click ‘Train Leo’ and enter COVID-19 as the topic.

Preview the prioritized COVID-19 articles in your Threat Research feed

You can see the most recent coronavirus-related publications from your sources in the preview. Choose whether to filter by Entire Content or titles that explicitly contain COVID-19 or its aliases.

New threats to your business can then be spotted and prepared for.

COVID-19 and retail

You’re a business intelligence analyst searching for COVID-19’s effects on stores and brands around the globe. Retail, one of the most disrupted sectors, is under intense scrutiny. The prioritization feature can help here too. 

With a Retail feed, you’ll preview countless pieces of content that tackle this subject. 

Again, just create a Leo priority around COVID-19.

Preview the prioritized COVID-19 articles in your Retail feed

And that’s it. You have a feed at the intersection of two subjects, with plenty of room for more priorities and further refinement.

Muting COVID-19

You might want to look past COVID-19 instead, and keep it out of your feeds. 

Muting is just as easy. Click ‘Train Leo’ and scroll to ‘Mute Filters’. Type in COVID-19. You’ll see a message asking which Feedly feeds you want to remove it from.

Here’s how it looks in a Tech feed. 

Preview the muted COVID-19 articles in your Tech feed

No more content on the topic will turn up in your Feedly, as long as the mute is active. It’s one of 1,000 pre-trained topics that Leo can mute right away.

Train Leo to prioritize or mute COVID-19 now

Whatever happens with coronavirus and your market, the trusted insights are here. Leo makes sure you’re never overwhelmed or struggling to see the big picture.

If you’re interested in learning more about Leo’s roadmap, join the Feedly Community Slack channel. 2020 will be a challenging year, but by staying informed, you can respond better and remain in control.

Mute Market Reports with Leo

We heard from lots of users that market reports can be a considerable source of noise when you use keyword alerts to track updates about companies.

We are excited to announce a new Market Reports Leo topic. We have taught Leo to read articles and understand if they are about market reports so that you can easily mute them from your feeds and save hours.

Let us show you how it works

Let’s imagine you have keyword alerts to track updates about various Health companies such as Amgen, Novartis, and 23&Me.

Market reports represent a large portion of the articles in our feed

As you can see, a considerable amount of these articles are noisy market reports. Let’s train Leo to read this feed and filter out all the market report articles.

You can find the Mute Filters skill when clicking on Train Leo.

Find the Mute filters skill in your feed

In the Mute Filters editor, you can select the topics and keywords you want Leo to mute. Search for the new Market Reports Topic.

Search for the new #market reports Leo topic

You can see a preview of all the articles that Leo has recognized as Market Reports and that will be removed from the feed.

Leo mutes articles he recognizes as market reports

Leo will continuously read your feed and remove articles he identifies as market reports, letting you focus on the topics and trends that matter to you.

Our feed is now free from any noise coming from market reports

The Leo Market Report Mute Filter helps us cut through the noise and track company updates a lot more efficiently.

Yuan Shen Yu

Train Your Leo Now

We are excited to see how many Feedly users declutter their feeds and dig deeper into the topics and trends that matter to them. Sign up today and discover what Leo can do for you!

If you are interested in learning more about Leo’s roadmap, you can join the Feedly Community Slack. 2020 will be a thrilling year with new skills and bold experiments!

Leo and Summarization

Reading through a large number of articles every day can be time-consuming, especially if those articles are long.

Helping you save time is a problem we are very passionate about, so we are excited to release today a new Leo skill called Summarization.

We have taught Leo to read and summarize the articles in your feeds so that you can more efficiently scan through articles and determine which ones are relevant.

Demo

Leo automatically reads all the articles in your feeds and summarizes them.

Articles lists showcase those Leo summaries as articles descriptions

Leo summaries in article lists

When you open an article, Leo also highlights the key sentences which are part of the summary. The goal is to help you get to the key insights more efficiently.

Leo reads and highlights the most important sentences

Board newsletters and slack integration also take advantage of the Leo summaries for the article descriptions.

Available Now

The Leo Summarization skill is available now to all users in the Pro+ and Business plans.

If you prefer not to see the blue highlights, you can turn them off via the Leo Summary Highlights preference.

If you have feedback about the Leo Summarization skill, you are welcome to join the Feedly Lab slack channel and discuss it with the product team.

Leo and Topics

Broad business and tech publications produce hundreds of articles per week. Not all those articles are relevant to the topics, companies, or products you care about. Manually filtering out the noise can be overwhelming and time-consuming.

Relevance is a problem we are very passionate about. We have spent the last two years designing and building Leo, your AI research assistant to help declutter your feeds and save time.

Unlike black-box recommendation engines, Leo has a set of skills that let you define and control what is relevant to you.

We are excited to show you how the Leo Topic skill lets you track specific topics, companies, and keywords in your feeds.

Let’s get started!

Companies, People, and Products

Leo knows about all the companies, people, and products listed in Wikipedia and in the news. You can ask Leo to look for any of those named entities (and their known aliases) and prioritize articles that are a match.

You can, for example, look for mentions of your competitors or prospects in your industry or tech feeds.

Train Leo to prioritize mentions of Tesla across a set of trusted business sources

Smart Topics

Leo understands how to recognize articles about hundreds of “smart” topics (like artificial intelligence, cybersecurity, blockchain, energy, health, etc..). He’ll be looking for thousands of different terms related to that smart topic. We designed smart topics because an article can be about artificial intelligence without including the term “artificial intelligence”.

Train Leo to prioritize #AI across a set of broad business sources

We continuously teach Leo new smart topics. If there is a specific topic you would like to sponsor, please email leo@feedly.com

Keyword Matches

You can also ask Leo to look for exact matches of a keyword you are interested in. In this mode, Leo behaves like a saved search.

Train Leo to look for exact matches of the “downsizing” keyword in your business feeds

Composable with AND and OR

You can design more sophisticated priorities by combining multiple topics using AND and OR. AND means that both of the topics need to be present. OR means that either of the topics needs to be present.

Train Leo to look for mentions of DNA or CRISPER and cancer in your health industry feeds

Composable with Other Skills

The topic skill can be composed with all the other Leo skills allowing you, for example, to easily prioritize articles that reference a product launch (business event skill) while also being related to #artificial intelligence (topic skill)

Train Leo to prioritize product launch articles related to #AI

Or high severity software vulnerabilities (cybersecurity skill) related to docker (topic skill)

Train Leo to prioritize critical Docker vulnerabilities

Continuously Learning

You can use the Leo “less like this” down arrow to correct Leo when a topic detection is incorrect. This feedback is channeled to the Feedly ML Team and to the datasets used to train Leo, making topics increasingly more accurate and relevant over time.

Leo continuously learns from your feedback

Available Now

The Leo Topic skill is available now on both Web and Mobile for both Feedly Business and Feedly Pro+ users. We look forward to seeing how you train your Leo and how much time you can save.

If you have feedback about the Leo topic skill, you are welcome to join the Feedly Lab slack channel and discuss it with the product team.

Leo understands funding events, product launches, and partnership announcements

Industries are changing at a faster pace than ever. Keeping up with new threats and opportunities can be overwhelming and time consuming.

Today, we are excited to announce a new Leo skill that lets you easily track key business events like funding events, product launches, or partnerships.

Here is a quick demo

Funding Events

We have trained Leo to detect and understand funding events. This means that you can now ask Leo to read your tech, business or industry specific feed and prioritize articles related to funding events – saving you a tremendous amount of time.

Track funding events in your feeds

Product Launches

We have also trained Leo to detect and understand product launches.

Track product launch announcements in your feeds

This means that if you are part of a sales or sales enablement team, you can ask Leo to read TechCrunch or New York Times and notify you each time one of your customers or prospects launches a product. You can leverage that product launch event to create a warm approach and engage in a smart conversation.

Partnership Announcements

Finally, you can also easily prioritize the fraction of articles referencing partnership announcements.

Track partnership announcements in your feeds

Composable with other skills

The business event skill can be composed with all the other Leo skills allowing you, for example, to easily prioritize articles referencing a product launch (business event skill) and related to #artificial intelligence (topic skill).

Track funding announcements related to #artificial intelligence

Trained across 24 industries

Different industries use different vocabulary to describe these business events so we trained Leo across 24 different industries.

Leo’s industries

Continuously learning

You can use the Leo prompt or the “less like this” down arrow to correct Leo when the event detection is incorrect. This feedback is channeled to the Feedly ML team and to the datasets used to train Leo.

Tell Leo when he has detected a wrong event so that he can learn

Available now

The Leo business event skill is available on Web and Mobile now for Feedly Teams users. Because we have been getting a lot of Leo requests from non-Team users, we will also be launching a $12/month Pro+ plan in October which will include Leo, a Twitter integration, and all other Pro features.

If you have feedback about the Leo business event skill, you are welcome to join the Feedly Lab slack channel.

Leo and Mute Filters

Some of the sources you follow in Feedly are broader than the topics and trends you care about. That additional noise can add up and become overwhelming or result in you wasting precious time.

We believe that noise is the enemy and we have been building a new Leo skill called Mute Filters to let you cancel that noise.

In this article, we will show you how to use Leo mute filters to mute companies, people, topics, authors, sites, and more.

Let’s get started!

Mute companies

Curating content to share on Social Media and want to avoid mentions of your competitors?

Train Leo to mute mentions of SAP in your business feed

You can train Leo to mute each of your competitors and automatically remove all the articles mentioning those competitors.

Mute people

Want to avoid articles about specific celebrities, politicians, or executives?

Train Leo to mute mentions of Kim Kardashian

Creating a Leo mute filter for a celebrity, politician or executive will automatically remove all the articles that mention that person from your feed.

Mute keywords

Want to avoid a spoiler about Game of Thrones until you have finished reading all the books or tired of hearing about Pokemon Go or the latest Apple monitor?

Train Leo to mute Game of Thrones

You can train Leo to mute specific keywords and remove all mentions of those keywords from your feeds, temporarily or permanently.

Note: with Leo Mute Filters, you no longer need to use quotes around phrases with spaces. Leo will take care of converting the input into the right query.

Mute topics

Following a broad source like TechCrunch, Wired and Forbes but do not care about topics like gaming? Or following a keyword alert for a public company but do not care about financials or market reports?

With Leo mute filters, you can mute topics and increase the focus of your feeds. Leo ships with 1,000 pre-trained topics.

Mute authors

Do not like a specific author from one of the sources you follow?

Train Leo to mute a specific author with the author: operator

With the author: operator, you can train Leo to look for specific authors and mute all the articles from that author in your feed. (Sorry Katherine, we actually love your work!)

Mute title patterns

Want to remove articles which have a specific keyword in their title?

Train Leo to look for a keyword in the title of an article

With the title: prefix, you can train Leo to look for a mention of a keyword in the title of the article and mute the matches.

Mute sites

Finding some of the sources referenced in Google News Keyword Alerts irrelevant?

Train Leo to mute specific sites using the site: operator

With the site: prefix, you can train Leo to mute specific sites from your keyword alerts.

Forever or temporarily

When you create a Leo mute filter, you can specify a duration.

Select a duration

Once you have trained Leo with a mute filter, you can easily remove, pause or resume that priority via the Train Leo page.

Pause or remove a mute filter<br>

Like with all the other Leo skills, it was important for us that you always feel in control and can continuously refine your Leo as your needs evolve.

While reading

When reading articles, Leo will highlight the most salient entities mentioned in the content. This makes it easy to click on them and priorities or mute those entities.

Mute an entity while reading

You can also highlight any snippet of text and mute that phrase

Highlight and mute any phrase

Finally, when reading an article, you can click on the Less Like This button and easily mute one of the topics Leo has associated with the article

Train Leo to mute a topic vis Less Like This

On mobile or on the web

The Leo mute filter skill is available both on the Web and on mobile (version 65+).

You can train Leo to mute topics and keywords on mobile.

From a feed

Train Leo to mute mentions of Apple on your Business feed

From an article

Train Leo to mute mentions of Spark New Zealand

From less like this (long swipe from right to left)

Train Leo to mute a topic via Less Like This

Curious about trying Leo Mute Filters on some of your feeds? Join the Leo program

FAQ

What happens to mute filter v1?

Pro users will be able to continue to use a more basic version of mute filters. The syntax of those mute filters have changed to the v2 syntax to allow more efficient processing on the back end.

Some of the v1 mute filters using advanced queries can not be migrated to v2 will remain active as legacy filters until user delete them.

Are there limits to the number of Leo mute filters a user or team can create?

One of the benefit of the Leo mute filters is that they can be processed more efficiently by our back-end. As a result, we are increasing the limit of Leo mute filters for Teams user from 100 total to 100 per feed.

Can non-Teams user access Leo?

We will be offering Leo to non-team users later this year via a Feedly Pro+ priced at $12/month. You can request early access to Pro+ here.

Can a mute filter target a specific source?

No. Mute filters can target a list of sources (what we call a feed) or all your feeds.

Leo understands Vulnerability Threats

Do you need to keep up with the latest vulnerabilities and threats but do not have the time to read all your security feeds? We can help.

In 2018, fifteen thousand vulnerabilities were discovered, the number of exploits doubled and more than four security articles were published every minute. Keeping up with all these trends can be time-consuming and overwhelming.

This is a problem we are very passionate about and have been researching with two of the largest security teams in Silicon Valley.

Today, we are excited to announce a new Leo skill called Security Threats.

We have been teaching Leo to read security articles and find or assess the severity of the software vulnerabilities they mention so that he can help you focus your attention on the most critical threats in your feeds first.

Here is a demo!

Let’s look at how you can train your Leo to prioritize articles mentioning critical vulnerabilities related to Microsoft, WordPress, or Docker.

Cut through the noise

Leo reads and prioritizes the most critical threats in your feeds

Leo continuously reads your feeds and short-lists the most critical vulnerabilities in the priority tab.

For example, you might have a cybersecurity feed connected to niche security experts, vulnerability databases, keyword alerts, etc. with thousands of new articles per month.

You can train Leo to read those 1,000+ articles and prioritize the 30 or so referencing high severity threats (CVSS > 8) and related to vendors you care about (Microsoft, WordPress, Docker in the example above).

Leo’s new Security Threat skill

You’re in control

Leo is not an opaque recommendation engine. Instead, Leo has a set of skills that gives you control over defining what information is important to you.

The new Security Threat skill allows Leo to read an article, lookup CVE, CVSS, and exploit information from multiple open source databases and determine how critical a vulnerability is.

The new Security Threat skill also includes a sophisticated machine learning model that allows Leo to assess the severity of a threat based on the vocabulary used to describe the software vulnerability. This is particularly useful for zero-day vulnerabilities which might not have a CVE or CVSS.

Training Leo to prioritize vulnerabilities is very simple.

Creating a Leo cybersecurity model

The first layer of the model captures the severity threshold. High means CVSS > 8 or CVSS > 5 but with an exploit.

The second layer of the model captures the list of vendors.

Control and transparency are core Leo design principles.

All the articles prioritized by Leo have a green priority marker. Clicking on that marker offers an explanation of why the article was prioritized and the opportunity to refine, pause or remove that priority.

Full control and transparency

When an article is related to a CVE, you can also click on that CVE to get additional information about the vulnerability: description, CVSS score, exploits, patches, etc.

Quick access to CVE information

Continuously learning and getting smarter

Leo learns from his mistakes. When a recommendation is wrong, you can use the “Less-Like-This” down arrow button to correct Leo.

Leo learns from Less Like This feedback

You can let Leo know that he misclassified a vulnerability, miscalculated the severity, or misidentified a vendor.

Leo learns from your feedback and gets continuously smarter.

Streamline your open-source intelligence

We are excited to see many security teams declutter their feeds and dig deeper into the vulnerabilities that matter to them. Sign up today and discover what Feedly for Cybersecurity can do for you!

If you are interested in learning more about Leo’s roadmap, you can join the Feedly Community Slack. 2020 will be a thrilling year with new skills and bold experiments!

What is new in Leo 0.6?

We pushed Leo 0.5 to a limited beta in early March and collected lots of interesting feedback. The team is listening and crunching through all that feedback and adapting Leo to improve UI/UX as well as the relevance of the underlying machine learning models.

Here is a summary of the changes we are pushing out today as part of Leo 0.6 Beta

Smart Topics

One of the feedback we collected was that the difference between mentions and topics was not clear. So in 0.6, we merged these two concepts into a single one we call Smart Topics. Just search what you want to prioritize and Leo will start analyzing the content of your feeds and prioritize the articles which are a match.

Search for companies, products, people and topics in a unified experience

Level of Aboutness

Sometimes you are interested in a company, product, or topic and you want to see every article mentioning that topic. Sometimes, for more popular topics, you are only interested in reading an article if the article is truly about that topic or company.

Leo 0.6 exposes a “level of aboutness” knob that gives you more control over the model so that you can cut out low salience matches.

Tune the aboutness parameter of each layer

For example, if you are interested in NLP or BERT, you can train Leo to only prioritize research articles that are prominently about those topics (as opposed to articles which only briefly touch on those topics).

This is a particularly powerful feature when combined with Google News Keyword alerts.

Global Priorities

Some Leo 0.5 beta customers mentioned that it was critical for them to be able to define priorities that span across multiple feeds. For example, you might be doing research about Stablecoin and want to prioritize that topic across both your Tech feed, your Business feed, or all your personal or team feeds.

In Leo 0.6, the priority designer allows you to pick “All Team Feeds” or “All Personal Feeds” as the scope of the priority.

Create a priority that spans across all your team feeds

This change reduces the total number of priorities you need to create and manage when researching topic and trends across multiple of your feeds.

Quick Access

Some users mentioned that they would like to be able to navigate their content by priority. If you are interested in a specific topic like Docker, it makes sense to be able to quickly see if there are new Docker related articles in your Feedly and easily access those articles.

In Leo 0.6, we added a new Priorities section to the left navigation bar that surfaces all your priorities and gives you quick access to all the article Leo has flagged as important.

Quick Access to all the NLP article prioritized by Leo

We added two settings in the Leo settings to let you personalize this feature. You can decide if you want to see priorities in your left navigation. If you want to see all the priorities or all the global ones (default). If you want to see all the priorities or only the ones which hav unread articles.

Inlined Entities

Your interests and priorities are continuously evolving. Often, you discover a new company, product, or topic while reading an article and you want to be able to teach Leo about it.

In Leo 0.6, the most prominent topics mentioned in an article are highlighted so that you can quickly prioritize them (or mute them)

Inlined Entities allow for quick prioritization of new topics

As part of Leo’s Cyber Security skill, you will also see highlights of CVE entities. More to come soon.

Like for the Quick Access feature, there is a Leo setting that allow you to turn off Inlined Entities if that is your preference.

Like Board Improvements

The ML team is spending time understanding how you are engaging with your priority feeds (which articles are saved to a board, which articles are being Less Like This’ed) and tuning the underlying ML models to improve accuracy. You should expect to see the quality of your priority feeds improve over the next 8 weeks.

Power Search

A lot of Feedly Pro and Feedly Teams customer rely on power search to find specific articles in their feeds and boards. In Leo 0.6, we are expanding power search and let you search with your priority feeds.

Search for BERT within the NLP priority

For teams using Leo to discover and track trends, opportunities, and trends across industries, the combination of Leo priorities and Power search is a powerful way to quick find the most crucial information

Thank you!

We want to thank all the beta customers who have been working very closely with us over the last few weeks (and sometimes months). We are very grateful for your time and precious feedback. This open collaboration is not only powerful and efficient but it is also very fun. We look forward to the next 3 months!

Edwin, Remi, and Victoria

Love reading? Love the Web? Join the Leo Beta Program

Introduction to Leo 0.5

Sometimes you want to follow high volume publications like The Verge, NY Times, or VentureBeat because you trust them but you are only interested in narrower topics, trends, or mentions.

Reducing noise and information overload is a problem we care passionately about. We have been working over the last 12 months on a new feature called Leo. You can think of Leo as your non-black-box research assistant – an easy-to-control AI tool which helps you reduce noise in your feeds and never miss important articles.

Here is a quick overview of the Leo 0.5 Beta feature set.

New Priority Tab

If you are part of the Leo 0.5 Beta Program, each of your feeds has now 2 tabs.

Introducing the new Priority Tab

The All Tab includes all the articles published by the sources you follow.

The new Priority Tab includes the subset of articles flag by Leo as important – based on the priorities you defined for your Leo.

Three Core Prioritization Skills: Mentions, Topics, and Like Board

Leo 0.5 ships with three core skills: mentions, topics, and like-board. Each of these skills allow you to prioritize articles differently.

The mentions skill allows you to prioritize articles based on mentions of people, company or keywords which are important to you.

Ask Leo to prioritize articles mentioning JP Morgan

For example, you can ask Leo to prioritize all the articles that mention “JP Morgan”

The topic skill allow you to prioritize articles which are about a specific topic you are interested.

Ask Leo to prioritize articles about quantum computing

For example, you can ask Leo to analyze your tech feed and prioritize articles which are about artificial intelligence, quantum computing, or gaming.

Leo ships with one thousand pre-trained topics. If the topic you are interested in is part of that list, the topic skill is a powerful tool to let you focus your feed on what really matters to you.

Sometimes, the topic you are interested in a very niche. This is where the Like Board skill is very useful and powerful.

Prioritize articles similar to the ones saved in your Smart Venue board

For example, if you are in the Sports industry, you might be interested in the emerging Smart Venue trend. Leo does not know out of the box about Smart Venue but if you can create a board and save 30-50 articles about Smart Venue, you can use the Like Board skill to teach your Leo a new personalized topic and ask Leo to prioritize future articles which are similar to the ones you save in that board.

Once you have defined the priorities of your Leo, he will continuously read your feed and flag articles which are aligned with those priorities.

The Like Board is particularly powerful because the more articles you save to that board, the more accurate Leo’s recommendation will become.

Finally, you can easily define more sophisticated priorities by combining multiple skills/layers.

Combine multiple layers

Feedback Loop Via Less Like This

When Leo makes a back prioritization, you have the control to provide him feedback via the Less Like This button.

Provide Leo feedback via Less Like This

There are 5 different classes of feedback you can offer to your Leo:

  1. The “Not About” feedback allows you to teach Leo that it matched the wrong keyword or topic. For example, you were interested in ICO (Initial Coin Offering) and Leo detected ICO (Internet Commissioner Office).
  2. The “duplicated article” feedback allow you to flag articles which are on topic but you have already read about via a different source
  3. The “I’m not interested in” feedback allow you to flag class of articles you are not interested about. For example, you might not be interested in market research type articles. If you can flag 10-20 articles as I am not interested in market research, Leo is going to learn and start prioritizing fewer market research articles.
  4. Sometimes (specially for keyword alerts), you might get articles from sources you do not care about. The ‘mute domain’ feedback allows you to train your Leo to mute articles from those domains.
  5. Finally, sometimes, the reason is more complex. The ‘Something else’ feedback offers you an easy way out.

Control and Transparency

A very important aspect of the Leo promise is that it is a fun, non-black-box AI you fully control and can easily collaborate with.

Transparency via clear explanations

Transparent because each time Leo makes a prioritization, he will explain why the article was prioritized and give you the opportunity to refine that prioritization.

Full control

Control because you explicitly define all the priorities of your Leo and you can at anytime go in the Train Leo section and remove or refine a priority. No black box. No lag.

Goodbye Information Overload

Leo 0.1 Alpha customers saw 40-70% noise reduction on their feeds. More targeted feeds mean that you can save time while reducing the risk of missing important articles, or being the last to know about an important risk or market opportunity.

We look forward to seeing how your will be training your Leo!

-Edwin, Remi, and Victoria

Deduplication Skill – Leo

It is frustrating to be skimming through your feeds and run into duplicate articles.

This happens for example when you have overlapping keyword alerts where two different keywords exist in the same article. It also happens when some sources publish the same articles into different RSS feeds. Finally, it happens a lot when a company issues a press release and other sources publish that press release with some minor changes.

Giving you the tools and control to tune your feeds is something we care passionately about. Today, we are excited to announce the beta release of a new Leo skill called Deduplication.

What is Deduplication?

This skill helps Leo detect that multiple articles are near exact duplicates of each other and cut that noise from your feeds. On the Web version of Feedly, you will see a small notification at the bottom right of your screen each time Leo removes duplicate from your feeds.

Which language does Deduplication work on?

The Leo deduplication skill works across all languages?

Which Feedly Plan does this skill require?

Because processing duplicates at scale is expensive, this skill will be initially rolled out as part of the Feedly Teams plan.

If you are part of Feedly Teams, there is a preference knob in the Leo settings page to disable this skill.

Beyond near exact duplicates

The deduplication skill is focusing on near exact duplicates. These are articles which have 85% or more overlap. We are working on a different skill called Business Events for articles which are reporting on the same event but with different content. In the case of business events, the content will be grouped instead of being removed.

Thank you!

We want to thank you Aymeric Bernard and Iheb Benabdallah for doing the preliminary ML research behind this Leo skill!

How to do NLP and machine learning with Feedly

Are you passionate about the Web, reading, and NLP? Are you curious about how machine learning can help process, filter, and prioritize information more effectively?

The goal of this tutorial is to show you how to leverage the content of your Feedly feeds and boards to run machine learning and NLP experiments using the Feedly API and Colaboratory.

To make things concrete, we are going to create a simple KNN classifier that takes as input the content of The Verge and Engadget, and a board with 50 positive AI articles. Using these inputs, the classifier reliably learns to classify AI and non-AI articles from those sources.

The Colaboratory Cloud Notebook we created has all the building blocks you need to create and run this AI experiment. All in a browser. All this within 20 minutes.

Once you are done with this first experiment, you will also have an example you can easily adapt to your own feeds, boards, and machine learning models!

Open Notebook

-Quentin

Passionate about the Web, NLP and Machine Learning? Join the Feedly Lab on Slack and connect with the Feedly machine learning team!

Topic Classification Skill – Leo

Sometimes you want to follow high volume publications like The Verge, NY Times, or VentureBeat because you trust them but you are only interested in narrower topics like #AI, #MixedReality, or #Blockchain.

Reducing noise and information overload is a problem we care passionately about. We have been working over the last 12 months on a new feature called Leo. You can think of Leo as your non-black-box research assistant – an easy-to-control AI tool which helps you reduce noise in your feeds and never miss important articles.

At its core, Leo is composed of a set of NLP and ML skills that allow you to train him to read, understand, and prioritize the content of your feeds. One of the first skills we are building is Topic Classification.

Leo Topic Modeling Skill

What is Topic Classification?

This skill helps Leo classify articles across 1,000 common topics. Thanks to its topic classification skill, Leo will be able to determine that the article above is about the #retail and #china topics.

Why is this skill useful?

Some sources produce interesting and high-quality content across lots of different topics. Following those sources can result in a large number of articles being added to your feeds, some of which about topics you do not care about. Let’s say that you like Techcrunch, Forbes, and Wired. Following those three sources means that you have to crunch through a thousand articles per week. Some of them on topics you do not care about. With the Leo topic skill, you can train Leo to prioritize articles about #ai across those sources and get the ten articles which are the most relevant to you without having to skim through a large list of articles.

How is Leo learning?

We feed Leo with thousands of articles about each of the thousands of topics that seem to be the most relevant to the Feedly community and generate a terminology and topic model that best captures the nuances across these topics.

Leo then processes all the articles in your feeds and classifies them based on the topic model.

Which languages does Leo understand?

Right now our dataset is English only so Leo only classified articles written in English.

How can you participate?

Leo is in its infancy (Leo 0.1). We are going to need six to nine month of training and quality improvement for Leo to grow and become useful. As part of our Lab effort (more open and transparent development process), we are inviting the Feedly community to contribute to Leo’s training.

You might see at the bottom of some article, a Leo prompt asking you to train Leo about the topics associated with the article you just read. You can in a couple of clicks pick the topics which are relevant and the topics which are wrong.

If you are not interested in participating in the Leo program and think that the footer is a waste of space, there is a preference knob to turn it off in the Leo settings page.

Join the Leo Beta

Thank you

We want to thank Quentin Lhoest for doing the preliminary ML research behind this Leo skill!

Dig deeper

How to do NLP and machine learning with Feedly

Meet Leo, your AI Research Assistant

Keeping up with topics and trends you care about within a sea of articles can be overwhelming and time-consuming.

Filtering out the noise so you can focus on what really matters is a challenge we are deeply passionate about.

Today, we are delighted to announce Leo, your AI research assistant.

How Does Leo Work?

We have been teaching Leo how to read and analyze information so that he can declutter your feeds and deep-dive into topics and trends you care about. With Leo, instead of spending hours going through hundreds of articles every day, you can free your mind, focus on what matters, and stay on top of your field. 

Unlike opaque algorithms, Leo gives you total control over your feeds. Leo has a set of skills that help him understand the world and enable you to define what is relevant to you.

Leo allows you to prioritize topics, trends, and keywords of choice; deduplicate repetitive news; mute irrelevant information; summarize articles, and so much more. 

Leo’s skills help him read articles and understand the world

The Topic skill lets you prioritize specific keywords, mentions, topics, and trends.

The Like-Board skill lets you train Leo by example. If you have curated over the time a board of specific topics or trends, you can ask Leo to read that board, understand what you are interested in, and prioritize future articles he thinks you’re likely to save to that board.

The Business Events skill lets you track industry activities such as funding events, partnerships announcements, product launches, leadership change, etc. of your interest.

Leo is much more sophisticated than a simple news filtering tool. It’s a true AI that uses machine learning and NLP to filter out the noise.

Jon Henshaw (Lead SEO Analyst – CBS Interactive)

Let’s take an example!

Imagine that you follow a broad business feed connected to many sources with thousands of new articles per month.

You can ask Leo to read all of the articles and prioritize the most insightful ones in the new Priority Tab. This will save you an enormous amount of time.

Leo reads and prioritizes the most relevant articles in your feeds

With Leo, you are in control of the priorities.

Let’s imagine you are interested in the Facebook Cryptocurrency trend. With just a few clicks, you can train Leo on this new priority:

You are in control of the priorities

Once trained, Leo continuously reads all articles in your feed and prioritizes the ones mentioning Facebook and related to cryptocurrency.

Articles prioritized by Leo have a green priority label, which gives you a clear understanding of why the article was prioritized. You can then take further actions such as Refine Priority, Pause or Remove that priority.

Control and transparency

Leo is smart! He continuously learns from your feedback:

  • When you save an article to a board, Leo considers that action as a positive signal that reinforces Leo’s learning.
  • When Leo is wrong, you can use the “Less Like This” down arrow button to correct Leo and improve future recommendations.
Use the Less Like This down arrow button to correct Leo

Leo helps us to find the signals in the noise. With Leo, we can basically automate knowledge gathering and focus on growing our expertise.

Tino Klähne (Head of Strategic Design – Lufthansa Innovation Hub)

Train Your Leo Now

We are excited to see many Feedly users declutter their feeds and dig deeper into the topics and trends that matter to them. Sign up today and discover for yourself what Leo can do for you!

If you are interested in learning more about Leo’s roadmap, you can join the Feedly Community Slack. 2020 will be a thrilling year with new skills and bold experiments!

Named Entity Recognition Skill – Leo

Sometimes you want to follow publications like TechCrunch, The Verge, Forbes, Wired, … because of their high quality but you are only interested in articles mentioning a competitor, a product you are interested in, or a client you are trying to connect with.

Reducing noise and information overload is a problem we care passionately about. We have been working over the last 12 months on a new feature called Leo. You can think of Leo as your non-black-box research assistant – an easy-to-control AI tool which helps you reduce noise in your feeds and never miss important articles.

At its core, Leo is composed of a set of NLP and ML skills that allows you to train him to read, understand, and prioritize the content of your feeds. One of the first skills we are building is Named Entity Recognition (and Salience).

Leo Named Entity Recognition Skill

What is Named Entity Recognition?

This skill helps Leo detect people, companies, products in articles, map them to the right entity (disambiguation), and determine their salience (which entity is the focus of the article).

Why is this skill useful?

There are two interesting use cases related to the Named Entity Recognition skill.

The first one is around the power of disambiguation. Disambiguation allows you to prioritize ICO (Initial Coin Offering) and avoid seeing articles related to ICO (Information Commissioner’s Office). When Leo sees ICO in an article, it will look at the context of the sentence and the article to determine if it is initial coin offering or information commissioner’s office and map the term to the right entity and refine its filtering. Apple the fruit versus Apple the company is another example of this.

The second is around the power of salience. Let’s imagine that you are doing some competitive watch on JP Morgan and you have created a keyword alert for JP Morgan. One of the limitations of keyword alerts is that they show you all the mentions of JP Morgan. So if you have an article about Amazon or Google and there is a short sentence about JP Morgan, you will see that article in your feed (even if that article is not about JP Morgan). With salience, you can train Leo to only see articles which are truly about JP Morgan (and remove all the irrelevant mentions)

How is Leo learning?

We feed Leo (deep learning model) with a set of article examples we have manually tagged with various entities. By reading through all these examples, Leo learns the structure of sentences and the occurrence of entities.

In a second module, we look at the extracted entities and try to map them to a knowledge base and disambiguate homonyms.

A third module reviews the list of entities across the article and determines the salience of each of the entities assigning an aboutness/importance score to each entity.

Which languages does Leo understand?

Right now our dataset is English only so Leo can perform named entity recognition on English articles only.

How can you participate?

Leo is in its infancy (Leo 0.1). We are going to need six to nine month of training and quality improvement for Leo to grow and become useful. As part of our Lab effort (more open and transparent development process), we are inviting the Feedly community to contribute to Leo’s training.

You might see at the bottom of some article, a Leo prompt asking you to train Leo about the entities (people, products, brands, etc.) associated with the article you just read. You can in a couple of clicks pick the entities which are the most relevant.

If you are not interested in participating in the Leo program and think that the footer is a waste of space, there is a preference knob to turn it off in the Leo settings page.

Join the Leo Beta

Dig deeper

Explanation of NER (Coursera video)

Hands-on: play with a pre-trained model

Hands-on: build a model in Stanford Deep Learning course

New AI-Driven Discovery Experience

We love the Web because it is an open and distributed network that offers everyone the freedom and control to publish and follow what matters to them.

We also love the web because it has enabled a new generation of content creators (Ben Thompson, Bruce Schneier, Tina Eisenberg, Seth Godin, Maria Popova, etc.). Those independent thinkers continuously explore the edge of the known and share insightful and inspiring ideas with their communities.

Connecting people to the best sources for the topics that matter to them has been core to our mission since the very start of Feedly.

But discovery is a hard problem. The web is organic, a reflection of the global community’s changing needs and priorities. There are millions of sources across thousands of topics and we all have a different appetite when it comes to feeding our minds.

About twelve months ago, we created a machine learning team to see if the latest progress in deep learning and natural language processing could help us crack this nut.

Today, we are excited to give you a preview of the result of that work with the release of the new discovery experience in the Feedly Lab app (Experience 06).

Two thousand topics

The first discovery challenge is to create a taxonomy of topics.

You can think of Feedly as a rich graph of people, topics, and sources. To build the right taxonomy, we started with the raw data on all of Feedly’s sources. We had to create a model to clean, enrich, and organize that data into a hierarchy of topics. Learn more about the data science behind this.

The result is a rich, interconnected network of two thousand English topics. And it’s mapped well with how people expect to explore and read on the Web.

Some topics are broad: tech, security, design, marketing. Some are very niche: augmented reality, malware, typography, or SEO.

On the discovery homepage, we showcase thirty topics based on popular industries, trends, skills, or passions. You can access all of the topics in Feedly via the search box.

The fifty most interesting sources

The second discovery challenge is to find the fifty most interesting sources someone researching any topic might want to follow.

Ranking sources is hard because not all sources are equal. In tech as an example, you have mainstream publications like The Verge or TechCrunch, expert voices like Ben Thompson, and lots of B-list noisy sources which don’t add much value.

In addition, for niche topics like virtual reality, some sources are specific to VR while others cover a range of related topics.

To solve this challenge, we created a model which looks at sources through three different lenses:

  • follower count
  • relevance (how focused is the source on the given topic)
  • engagement (a proxy for quality and attention)

The outcome is new search result cards. You can explore the fifty most interesting sources for a given topic and sort them using the lens that is most important to you.

Neighborhoods

One of the benefits of the new topic model is that the 2,000 topics are organized in a hierarchy. This makes it easy for you to zoom in or out and explore many different neighborhoods of the Web.

For example, from the cybersecurity topic, you can jump to a list of related topics that let you dig deeper into malware, forensics, or privacy.

One more thing…

We have done a lot of research over the last four years to understand how people discover new sources. One insight we learned is that people often co-read certain sources. For example, if you are interested in art, design, and pop culture and you follow Fubiz, there is a high chance that you also follow Designboom.

With that in mind, we spent some time creating a model that learns what sources are often co-read. The idea is that a user could enter a source that they love and discover another source they could pair it with.

You can learn more about the machine learning model (we call it feed2vec) powering this experience through the article Paul published here.

As a user, you can access this feature by searching in the discover page for a source you love to read. The result will be a list of sources which are often co-read with that source.

Thank you!

I would like to thank Paul, Michelle, Mathieu, and Aymeric for the great research work they did to take this project from zero to one. People who have tried to tackle discovery know that it is a very hard challenge and the results of this project have been very impressive.

We would also like to thank the community for participating in the Battle of the Sources experiment. Your input was key in helping us learn how to model the source ranking. We are going to continue to invest in discovery and we look forward to continuing to collaborate with you.

We would also like to thank Dan Newman, Daron Brewood, Enrico, Joey, Lior, Paul Adams, Ryan Murphy, and Joseph Thornley from the Lab for reviewing an earlier version of this article.