— Export your bookmarks, bulk actions, repository overview and more…
We’re very excited to say we launched the latest version of our academic tools this week, version 6.1! As always, we’ve improved the backend and done some bug fixes to ensure you have a smooth experience using Iris.ai — but we have of course some juicy features for you as well.
I’ve listed the new features below. Do you want to try them out? Register and sign in to get started with a free account. We hope you enjoy the new update! ????
Now you can finally export all the papers you have bookmarked. Simply, click the brain icon in the top right corner, go to your reading list and select the papers you’d like to export. You can choose between CSV and BIBTEX.
Bulk actions
We’ve added checkboxes in front of every entry, allowing you to bulk actions. For example, exporting multiple papers.
Free or premium?
If you were ever in doubt, now you can see whether you’re subscribing to a premium or free version of Iris.ai — whether you’re an individual subscriber or a university member.
“But where are all these papers coming from?”
One question our users often ask us is, “but where are all these papers coming from?” Now we’ve given you a neat way to quickly check the repositories in which the papers were found. (Also, hot tip: if you find repositories in your list that you don’t want to see papers from, use the repository filter to exclude them!)
Undo/redo actions in a session
When you’re editing your map in hierarchy, you’re now able to undo and redo all the actions you’ve done in that session.
Hierarchy Editor: Duplication of concepts
Using the hierarchy editor, there are sometimes subconcepts that are highly relevant under multiple top concepts. Now we’ve given you the ability to duplicate the concepts so you can place them in several parts of the fingerprint — more customization and control over the fingerprinting process.
If you have any features on your wish list that haven’t been created yet, let us know at support@iris.ai!
We are a company that do our best to put impact first. Over the last few weeks we have found ourselves initially stunned and perplex, and eventually a tad bit overwhelmed, by the COVID-19 pandemic and the SARS-CoV-2 virus causing it. What can a small startup team do, beyond sharing our advice on remote work (we’ve been remote since the start of the company), do our best to keep our employees safe and sane and stay hopeful that our revenues will not decline entirely to zero?
We can’t do much. But if you are a researcher, medical professional or problem solver, you can. And we can help.
For anyone working on any aspect of research around COVID-19, whether on the epidemiology, virology, biology, psychology or anything else – and whether you are an academic or a concerned citizen – we hereby invite you to join Iris.ai, and your account will then be added to the premium access COVID-19 research group. Our only requirements are that your work is on the current pandemic, and that you allow your findings through Iris.ai to be shared publicly.
Our tools can help you find the right research for the problem you face. Here, for example, is an interesting map on “The epidemiology and pathogenesis of coronavirus disease (COVID-19) outbreak”.
The amazing humans at the Allen Institute / Semantic Scholar have worked on an open dataset of almost 30,000 articles related to COVID-19. This, plus about 120 million other research papers (Open Access), are all connected to the Iris.ai tools.
On April 7th, at 8am CET and 5pm CET we will host free webinars for everyone joining the premium COVID-19 research group (register here), to help you get started, answer all your questions, and do anything we can to help your research process run smoothly. Please sign up to get the details.
Everyone is of course also welcome to join our monthly online workshops that walk you through the premium tools, whether or not you’re a COVID-19 researcher. Have a look at our Facebook events for upcoming dates.
In order to get set up with free access to the COVID-19 Iris.ai premium organization, here is what needs to happen: 1. You sign up for an account on https://the.iris.ai/auth/registration 2. Then send an email to order@iris.ai from the same email account and mention COVID-19 in the subject. 3. We will then add you to the premium group, and notify you when it’s done so you can log in, accept the terms and get working.
We will keep this premium access open to you for a minimum of three months – until end of June 2020 – and it will be extended unless the pandemic is over.
We don’t know what each one of us can do on our own, but we know how much power there is in a world coming together against a common enemy. We hope our tools might help you in the process. Let’s see if we can save some lives.
If you are not a researcher, but want to help, you can do so by sharing this announcement with the people in your network who might have a use for it.
Map subscriptions, BibTex exports, patents, dark mode and algorithm improvements – and some changes to free and premium versions.
It’s that time yet again – last night we pushed out some new and exciting changes to the academic tools! There’s a long feature list below, some of which (including map subscription and BibTex export) are long time favourite feature requests. We hope you enjoy them!
The main change with this launch is in the logic of free and premium accounts. We’ve made some choices on this front driven by a few reasons: a) We’ve now properly established ourselves with an offering toward university libraries, and found a collaboration model that universities can afford and that we can live with. That means more and more users are coming in through University licenses and have access to the full premium tools, which makes us very happy! b) We see that the decidedly best experience one can have with the tools is to have access to it all. c) At the same time, we wanted to give free users the ability to get a feeling for all of the tools, not just the Explore tool.
In short, as a free user we’re limiting some more of the features in the Explore tool, but opening up the ability to try out the Focus tool as well – and we’re starting to Beta test an individual payment model, if you want easy access to all of the tools and features! (Send an email to order@iris.ai for more information)
With that said, let’s dig into the details. The ability to use the tools without being registered even with a free account is heavily reduced. You can engage with shared Exploration maps and Focus studies, but you can not create them nor edit them.
For free user accounts, we chosen to limit some features including the ability to input your own problem statement and search in patents. You will see the Focus tool now being available in your account, and be able to import single Explore maps into a Focus study.
To see the full breakdown of free versus premium tools, go to iris.ai/accounts.
If you urgently miss some of the features we have had to remove from the Free tools, or want full access to the features of your new Focus tool, there are two ways to go about it: 1) Talk to your librarian about getting the entire university/department access to our tools or 2) Contact order@iris.ai to inquire about individual premium access.
New features
Dark mode: Let’s face it – we’re all different, and we figured we’d add a neat little feature for those of you who prefer the darker side of life. Toggle on dark mode on top of the left hand menu for a new Iris.ai experience.
Subscribe to maps: Have you built the perfect map that contains state of the art of what you need, and would love for us to let you know whenever new relevant content is found? It’s one of our most requested features and it’s here! Subscribe to your favourite maps and receive notifications whenever the content change with new relevant papers.
Easier content selection: Do you want to search by patents, papers, or both? This is now a lot easier to choose – either as you input your research question or starting paper, or in the map itself, which will regenerate the map.
Export your map results: Another highly cited request – this feature allows you to export the list of papers in your result map as a .csv file, or in BibTex format for uploading to your favourite reference manager system.
Bye-bye, TED talks! When we started building Iris.ai, we found a fun way to build something that made sense and demonstrated our intention and vision – but that didn’t require the full capability of reading millions of research papers. The concept we launched early 2016 was “see the science around a TED talk” and allowed you to pop a TED link into the tool, Iris.ai would machine read the script and show you a science map. Years later, while we will care deeply for our little hack, we figured it was time to retire support for this practically unused feature. So goodbye, little TED-Science-tool, it was a pleasure having you with us!
Algorithmic and capability improvements. We keep on improving the core engine, and the new improvements this time around are connected to an improved keyword extractor module (using tf-idf to propose candidates for key words, and then topic modeling to evaluate the relevance of those candidates) and the first part of better word disambiguation (w2v, currently disambiguating on part-of-speech, which helps finding better synonyms for e.g. words that are both noun and verb or other variations), and a new algorithm for hierarchy building to create smarter concept maps for the Explore tools (uses not only topic information and similarity, but also generality and concept dependencies to form the final output), freshly tuned hyperparameters that give better results overall – and some general system performance upgrades and updates.
Last week we announced the launch of Iris.ai 5.0 with improved algorithms for improved results. But what does that mean, exactly? And where did the Training platform go again?
Well, it’s all connected!
In this version of Iris.ai, we’ve improved Iris.ai’s computational power in a big way. When you click ‘Submit’ to create an exploration map, Iris.ai gets to work on generating a fingerprint of your input, containing the concepts, topics, synonyms, and hypernyms that could be used to describe the content. This, of course, is how it has always been.
In an ideal world, Iris.ai would be able to compare your input to the >100 million papers in her database within the 5-10 seconds it takes to generate a map. However, due to computational limits on earlier iterations of her algorithms, there was some amount of approximation that went into getting the most effective results in a reasonable amount of time.
Enter Iris.ai 5.0, with algorithms that allow over 20 times the computing power and thus can process with exact accuracy more of those 100 million papers. This results in less approximation and more relevant papers in your map. And, as we all know, more relevant results = better research, faster.
But, what does this all have to do with training?
Iris.ai’s learning is largely unsupervised but has used the training platform as a form of validation of her assumptions as an AI. With the help of our AI Trainer community, we were able to look at her assumptions against real, human inputs to ensure she was working as effectively as possible. With your help, we collected over 8,000 validation points against which to test Iris.ai. Thank you!
Of course, the training platform as it existed before today used questions and methods that validated the previous algorithms.
Now that new algorithms are in place, it is time to reassess the training platform to determine how our training community can assist in validating Iris.ai’s brand new algorithms. For the remainder of this year, we’ll be working on a revamped training system that will serve Iris.ai and our community of users and trainers even better as our little AI scientist grows.
Thank you all for your contributions as well as feedback and questions with each new iteration of Iris.ai. We love hearing from you! If you’d like to join the conversation or contact us directly, join the Iris.ai Facebook Group or contact community@iris.ai.
Haven’t tried Iris.ai 5.0 yet? Creating an account is free and easy! Get started here.
Iris.ai 5.0 is now live, and with a new version comes a number of new improvements to make Iris.ai smarter and friendlier than ever before.
First, the article Relevance score features more prominently in the Explore tool. Now, users will notice the relevance calculation right on each map cell, along with other high-level article information.
And, marking an article as “Relevant” will color code that cell for quick reference later on. Those marked as not relevant will now be removed from the map.
Finally, the filters section now features a Relevance filter, with the ability to set a relevance threshold so that you can broaden or narrow the scope of your results based on how closely each paper is related to your query.
Of course, all of this is fabulous news, but it’s made even better by the hard work our developers have put into honing Iris.ai’s algorithms to produce more relevant and accurate results than previous versions.
In-house testing revealed improvements in both volume and accuracy of results for the same query between version 4.3 and 5.0. Version 5.0 produced more than double the number of results, with a 13 percentage point increase in the number of resulting papers that were considered highly relevant to the original query.
Version 4.3
Version 5.0
169 Total Results
369 Results
128 High Relevance Results (75%)
324 High Relevance Results (88%)
41 Low Relevance Results (25%)
45 Low Relevance Results (12%)
Iris.ai is growing up quickly and still has her eyes on the title of A.I. Scientist. We hope these improvements will help you make new discoveries and speed up your next research project.
Interested in learning how your university or organization can benefit from the full suite of Iris.ai tools? E-mail our team to learn more and start a free trial of Iris.ai Premium.
Since our launch in 2015, over 10,000 brilliant community members have stepped up to become AI trainers, a volunteer team that helped Iris.ai to better understand over 9,000 different inputs. With each training session, Iris.ai was able to learn from human input, helping her to better understand language and serve better results.
As part of our upcoming release, the era of the AI training program closes, for now. With your help, we’ve made immense progress with Iris.ai’s human-assisted learning and it’s time to step back to review all of the learning from this program so that we can make improvements and determine what learning looks like for Iris.ai in 2019 and beyond. Thank you to everyone who participated over the years! We couldn’t have gotten here without you.
Of course, with one closed door, many more open! Stay tuned for upcoming news on the newest update for Iris.ai, including a number of exciting improvements to our most popular tool, exploration mapping.
With a new year, comes a new update to Iris.ai, with some big improvements we’re sure will help you fulfill your 2019 research goals.
Version 4.3 launched today and includes a number of back-end updates as well as a much-requested addition–the integration of PubMed into our searchable database of academic papers. This will add millions of papers to the possible results available through the Exploration tool.
Among the other updates, you’ll notice:
The ability to edit a problem statement after creating an exploration map.
Clarifying language in the Focus tool UI.
Improved keyword extraction for more relevant results.
Improved reliability and security.
Overall system performance improvements.
If you haven’t already, login and try them out for yourself! We would love to hear what you think of the improvements. We are always available for your feedback and questions. Just contact community@iris.ai.
Best wishes for a happy and productive new year, from the Iris.ai team!
It’s another big day in the life of Iris.ai, and we couldn’t be prouder parents, with the official launch of version 4.0. Last year brought our free users the option to create accounts, bookmark useful content, and revisit their browsing history. This time, we’ve introduced a new premium feature: The Focus Tool.
When conducting preliminary research on a topic, it’s easy to amass a list of hundreds, if not thousands of relevant research papers, especially with the help of Iris.ai. The Focus tool makes it easier to distill that list into a precise and—well, we’ll go ahead and say it—focused list of the most relevant articles to you.
While Iris.ai’s Exploration algorithm is great at surfacing a comprehensive landscape of interdisciplinary papers that relate to your original article or question, the Focus Tool allows the creation of intelligent filters to include or exclude topics of interest, retraining the algorithm as it goes.
This significantly reduces the average time it takes for professional researchers to compile a full report of relevant papers to support their work, in turn reducing a process that could take weeks to as little as two days and increases the confidence level of results by 15%.
Premium users will also find that they have a new way to search within the Exploration tool. You can now manually enter your research question and problem statement to generate a map, without the need to provide existing research on the topic.
Ready to see it in action?
Check out our latest video showing the new 4.0 tools and how they work together to get you a concise and manageable reading list in record time.
Interested in what premium can do for you or your organization? Visit https://iris.ai or e-mail maria@iris.ai for more information.
We’re happy to announce our baby Iris.ai has taken another little step – this time bringing our free users some new useful features we believe will make your lives a bit easier.
The obvious goodies – history and bookmarks in your own account
You know that feeling as you move through a map you’ve created, and you find just the paper you need – but just want to save it to read for later? We have you covered – bookmark it and you will find it again to download directly in the reading list in your dashboard. You can also save full maps, when you create a good one.
Researching a problem is a process, and we realize keeping track of where you’ve been and what you’ve seen isn’t always the easiest. So we do it for you; you can now visit your dashboard to review all the maps you’ve created and all the papers you’ve opened. This might come in handy if you suddenly remember something you saw earlier but didn’t realize at the time that it was relevant; we’ve now got you covered so you can retrace your steps from your dashboard.
There’s always more than meets the eye in these releases, and Iris.ai 3.0 doesn’t only come with new front end features but she’s gotten a solid algorithm upgrade too. New and improved data models, better tuning, some new neural network algorithms replacing older off-the-shelf components; we’re quite pleased with the results and while not as visible as the frontend we hope you will be, too.
Our new training tool is OUT. Starting today, the human friends of Iris.AI can train the AI directly from research paper abstracts and help her understand synonyms.
In May 2016 we launched the first version of our AI training tool to help Iris.AI learn via Ted Talks. A large labelled data set was needed to build an effective training loop. Since then we’ve asked our users to join us in our effort to take the existing stock of research into effective use by participating in this learning experiment.
Fast-forward 10 months and the crowd-training has become one of the backbones of our AI development. Iris.AI trainers around the world have done a tremendous job by training thousands of texts, equivalent to more than half a million trained concepts. These inputs have allowed us not just to improve the accuracy of our algorithm by several percentages but also to verify and assess the quality of it.
Today we’re taking the AI training to new heights by releasing the next version of our training tool. The new curriculum allows Iris.AI to learn directly from research papers and synonyms. This means that the source data of the supervised data set expands to millions of texts letting Iris.ai optimise her neural nets with scientific concepts across research fields.
Our next goal is to gather and inject a trained dataset of 5000 paper abstracts to the algorithm. With those inputs we aim to improve the connections in the neural nets of Iris.AI by approximately 10 %.
Interested in joining the effort and taking Iris.AI to the next grade? Sign up to become a trainer and we’ll help you get started.
Kudos to all our existing AI trainers and welcome new ones! We look forward to sharing the next leaps of this journey with you.