We at Iris.ai are on the edge of our seats with excitement with the upcoming launch of the Researcher Workspace! Why, you ask? The new tool suite is content focused, and allows you to follow your own workflow and processes. It can easily be adapted to your organization’s needs. The Researcher Workspace includes the modules: Search, Filter, Analyze, Extract, Summarize, Automate, Report. In this blog post we want to give you an overview of the new tool.
What differentiates the Researcher Workspace from our current tools is that there is no structured process that you must follow. You can upload your data from any source to your dashboard and decide what you want to do with it. Some of the ways you’ll be able to input information will range from using external URLs, cloud drives, open-source repositories to using files on your own device.
Once you have added documents that you want to investigate further, your main point of interaction with the data will be the dataset preview. It will allow for document preview so you know what you are playing with and the outputs from the other tools will be accessible in this one unifying place.
Search
Upon selecting at least one dataset, you will be able to search for literature by giving the system a self-written text or a link to a research paper – much like our existing Explore tool. As a result, you will get a list/map of relevant papers including their relevance scores. The results will automatically create a new dataset with an automatic context filter related to the input. This tool is great when you don’t really know what you are looking for, allowing you to do a broad search on a topic.
Analyze
After your dataset is created, you can use our word and topic analysis tool to filter through your results – similarly to our existing Focus tool, but more flexible. You can use generated words and topics as inclusion/exclusion criteria on your dataset so you can filter through it.
Filter
You can also filter your data using metadata filters for (1) publication dates and (2) add/remove repositories or (3) create filters using context descriptions based on your free-text. The context filter is saved automatically, so you can use it later on a different dataset. Context filters are especially useful when you are trying to filter something that cannot be described with one word but rather by using one or multiple descriptions.
Extract
You can as well automatically extract the data from selected documents into a tabular format, which has been pre-approved and made available by Iris. This feature allows the extraction of different types of data possibly corresponding to the different type of document selected for the extraction.
Summarize
The Iris.ai Workspace comes with a configurable summarization engine. You can quickly get summaries of the articles and check if there are any novel and relevant findings in them. So you can skim a lot of content faster. The engine can rapidly produce summaries of one or multiple abstracts or of one or multiple full text documents. The summarization that Iris.ai uses is abstractive which captures the context of summarized text. The generated summary contains new phrases and sentences that may not appear in the source text.
Report
If you want to document your search, we can prepare customized automatic report generation for you. The report can contain the number of searches, most used filters and more.
Every research process is a little different, and your Research Workspace will enable any workflow. In the future blog posts we will share with you more details, so stay tuned!
Medicines can cause certain unknown side effects that may emerge when they are on the market over a longer period of time. It is vital for pharmaceutical companies and regulators to constantly track such adverse events and report them. This is the fundamental principle behind a process called Pharmacovigilance, according to the WHO ”pharmacovigilance is the science and activities relating to the detection, assessment, understanding and prevention of adverse effects or any other medicine/vaccine-related problem”. At Iris.ai we have a set of smart, machine learning-based modules all based on machine understanding of scientific text and data, which is essential for Post Market Surveillance Automation.
According to recent reports the global pharmacovigilance market size is expected to reach USD 17.36 billion by 2030 and is expected to expand at a CAGR of 10.5% from 2022 to 2030.
Importance of Pharmacovigilance Post the Pandemic
Covid has provided various challenges as well as opportunities for pharmacovigilance service providers. During Coronavirus lockdowns there was a decline in ongoing clinical trials due to the restricted availability of clinical resources and research staff. The restriction of patients also brought changed work practices, routine audits and safety inspections. Additionally, an increase in the use of drugs (either approved for Covid or off label use) have driven the healthcare industry to work round the clock for safety profiling of the products. To be able to leverage such growth opportunities, the Pharmacovigilance industry is increasingly adapting to new trends attributed to better and more effective data collection and analysis. In this blog post we will share 4 main Pharmacovigilance trends for 2022.
Automation in Pharmacovigilance
There are many areas of pharmacovigilance where automation can help. As our customers have confirmed automating pharmacovigilance has many benefits – it eliminates human error, reduces costs and time. This helps manage large amounts of data and ensure compliance.
With our Iris.ai tools you can automatically find the documents you need by inserting any written text on the topic, such as product and disease descriptions, patents, case reports, clinical trials, etc. You can summarize them and extract all relevant information. Freeing up this time will enable PV professionals to spend time on more valuable-added tasks.
Pharmacovigilance requires basic automation to ensure automatic tracking, task monitoring and data collection. Automation can change the way data is collected and analyzed, which could accelerate clinical trials. Electronic data capture (EDC) is a database used to store patient data during clinical trials. The use of EDC-based tools for data collection and analysis during clinical trials and market observations is efficient. Cloud technology is ideal for providing a fully integrated database for all stakeholders, which is essential for improving drug safety and pharmacovigilance. Robotic Process Automation (RPA) eliminates manual tasks by providing automated data entry, processing and analysis. Combining RPA with cognitive automation through natural language processing can help with decision-making. Artificial intelligence and machine learning can help data analysts and data scientists make predictions based on data analysis. This can improve the quality of pharmacovigilance processes.
Scaling Resources
During the pandemic, there was a noticeable raise of Adverse Effects (AEs). To effectively manage this workload, the pharmaceutical industry had to scale their in-house resources and enhance their manpower. There has also been an increase in the manual tasks to be outsourced like data collection and entry. The scaling process has been improvised with adjustment of key performance indicators (KPIs) and dedicated teams, which has led to a greater degree of flexibility.
Dedicated Teams and Remote Work Culture
Building remote stations for a dedicated team of pharmacovigilance experts is a new and effective trend that drove the pharmaceutical industry during a pandemic. Such workplaces are divided according to job categories in order to achieve better coordination and specialization in the processes.
The importance of pharmacovigilance has been renewed and re-emphasized in the face of the global health crisis. Adapting to the new post pandemic trends will definitely increase cost effectiveness and improve the operational efficiency of the pharmacovigilance sector.
Importance of Collaboration
Collaboration and integration is the fourth trend for 2022. To improve, the healthcare industry, including pharmacovigilance, should close the gap between different sections from the medical affairs and safety areas to the clinical development and regulatory affairs space. Thanks to trends in the pharmaceutical industry, teams are more interconnected than ever before. For teams to be successful, they need to work more closely together, data must flow freely, and business processes need to be streamlined. Successfully implemented digital transformation strategies can have a significant impact on medical advances and patients’ lives. Our Iris.ai team is fully dedicated to facilitating our client´s pharmacovigilance processes to increase patient safety and drive innovation.
Key Takeaways
👉 The pandemic caused decline in ongoing clinical trials and a raise of Adverse Effects reported.
👉 Automation of pharmacovigilance process will drive the industry – EDC, RPA and AI are trending.
👉 Pharmacovigilance companies will scale their resources.
👉 Remote work culture is on the raise.
👉 Collaboration between different departments is crucial to improve the pharmacovigilance process.
Have you ever heard of the expression tl;dr? It stands for “Too long, didn’t read”. It likely originated on the comedy forum Something Awful around 2002, and spread to other online forums. Nowadays it is used by authors when they give a short overview of their text. Does it sound like something you might have said before too?
There are a lot of good examples for when you prefer to have a short summary of the document for instance if you are not sure yet if it is of interest for you or when you actually only need the essence of it to continue with your work.
A good example for that might be the book “On the Origin of Species” by Charles Darwin. Most likely all of us know his famous ideas on evolution on natural selection but only a handful of people have bothered to read that 502 pages long book. Simply because for us it has not been necessary to understand all the details but the main messages.
Article overload
The number of published papers and patents grows exponentially. About 2 million papers are published per year, that’s about 5,500 per day. And the growth is growing! So, to stay on top of all the new findings one has to navigate through a vast amount of literature. And after a long search for the right papers, one then has to read each paper and summarise it or at least understand the main takeaways. Instead of reading through lengthy papers; Wouldn’t you prefer to quickly understand the main messages of the paper and if it is interesting of course, go ahead and read it as your night time lecture?
If we can quickly assess the main findings and messages of one or a couple of similar papers, patents and internal documents, we would use less of our valuable time reading and summarizing papers. Hence, we can allocate our resources to more important tasks like using the empirical findings on inventing a new product or drug.
Iris.ai Summarization Tool
Our Summary Tool, as the name indicates, automatically produces summaries of any given document for you. In general, there are two types of summaries – extractive and abstractive. With an extractive summary one identifies the relevant information that is then extracted and grouped together to form a concise summary. In contrast to that, the Iris.ai Summary Tool creates abstractive summaries, meaning that the machine rewrites the entire document by building internal semantic representation, and then a summary is created using natural language processing.
The tool can make a summary of one as well as multiple documents. For the multi documents the machine identifies the parts of text in common and only makes a summary of that. Additionally, the summary can be based on abstract only or full text. The user can also adjust the length of the summary. We can take this feature a step further – for example, by you providing a topic of focus and the machine summarizing only text related to that, not the full texts. Let’s say you want a summary of only the experiments, or only the methods, or only one of the topics covered. This is still experimental but very exciting.
The tool can be integrated with your systems and the user interface can be customized.
Watch the full demo alongside with the peak behind the curtain of the technology we’re using here:
It’s almost the end of the year, which means time for predicting AI trends in 2022! There has been a dramatic growth of AI in the past few years. The acceleration of the adoption of digital technologies due to the pandemic has led to an increase in streamlining operations with connected technologies and automation. We can expect more AI trends to affect our daily lives in upcoming years.
Before we move onto the AI trends for 2022, here are some numbers. The AI/ML industry is expected to grow at a CAGR of 33% by 2027. It is estimated that businesses will have at least 35 AI initiatives in their business operations by 2022.
AI Metaverse
Metaverse (the prefix ‘meta’ and the word ‘universe’) is generally used to describe a virtual universe, accessed by virtual reality (VR) technology. It makes it possible for people to interact, do business, and build personas entirely online. The concept of the metaverse turned into a hot topic since Mark Zuckerberg rebranded Facebook into Meta.
Artificial Intelligence will be a vital point of the metaverse – from generating synthetic characters, to creating stories, to optimizing VR experiences. It will allow an enterprise to create a virtual world where the users will feel at home with virtual AI bots, which will assist them in picking the right products and services.
Low-code and No-code Technologies
The shortage of skilled AI developers or engineers slows down the adoption of AI technologies in many companies. No-code and low-code technologies are the solution.
To compare – currently there are web designing tools (for example WordPress) where users can just drag and drop the modules and features to the page and the website is ready. Similarly, no-code AI systems will help in creating smart applications by combining multiple pre-created modules and feeding them domain-specific data. NLP and Language Modelling could be used for giving voice-based instructions to execute various tasks. It will result in the democratization of AI, ML, and data technologies.
The augmented workforce
People are concerned that AI technologies will replace human workers and even make some jobs outdated. However, as businesses began to incorporate these technologies they found out that working alongside machines improves employees’ efficiency.
In engineering AI tools provide predictive maintenance. In marketing and sales AI is used to help determine which leads are worth pursuing and what value we can expect from potential customers. In knowledge industries AI is used to sort through the ever-growing amount of data that’s available to find the relevant information.
Iris.ai tools help researchers find more relevant scientific articles, extract information and summarize documents and researchers can save time up to 80%. You can learn more about our tools here.
Emerging technologies can help us do our jobs more efficiently, and it will be incorporated more and more in our daily working life in 2022.
Quantum AI
Companies will start using AI powered by quantum computing to solve complex business problems faster than traditional AI. Quantum AI can predict patterns and analyze data faster and more accurately. It can help businesses to identify unforeseen challenges and suggest possible solutions. As a result, Quantum AI will revolutionize many industrial sectors, such as healthcare, chemistry, and finance.
AI in cybersecurity
This year the World Economic Forum identified cybercrime as potentially posing a more significant risk to society than terrorism. As technology is more integrated in our daily life, the risk of cybercrime rises. The more devices you add to your network, it creates a potential point-of-failure that attackers can leverage to access data and misuse it.
AI can play a significant role in tackling this issue. For example, it can analyze higher network traffic and recognize patterns of nefarious virtual activities. We can expect a considerable amount of AI development in the area of cybersecurity in 2022.
AI in creativity
Creativity is widely considered a skill possessed only by humans. However, we are witnessing the emergence of creativity in machines. Artificial intelligence can create music, poetry, or even video games. With the upcoming launch of GPT-4 and Google Brain in 2022, we will refine new boundaries for AI in creativity. AI will be creating headlines for articles and newsletters, creating logos and infographics etc.
Key Takeaways
AI is developing dramatically and with it we should expect new areas and possibilities to use it. Companies should definitely consider using these opportunities to scale their existing business capabilities and gain competitive advantage.
Thanks to advances in technology, some computers are able to run business processes without human margin of error. Natural Language Processing (NLP) allows chatbots to understand speech and provide customer support in many industries. HR departments and finance companies use robotic process automation (RPA) to verify payroll systems and manage expenses. Computer vision is used to scan barcodes and track packages without the help of human hands. Artificial Intelligence is everywhere in our daily life.
The advancements in the technology and AI revolution cause people to question whether their job will be taken over. On the contrary, implementing AI solutions is actually creating demand for more knowledge workers. And even though some jobs will disappear from the market, other new ones will appear. Twenty, thirty years ago people were afraid that computers would take over their job – but here I am, working the job that wouldn’t exist without computers. A report from the World Economiс Forum shows that by 2025, 85 million jobs will be replaced by machines with AI, but 97 million new jobs will open by the same year.
Artificial Intelligence has been developed to the point that it can perform some easy and common tasks. AI has the potential to really disrupt the labour market. However, there are a number of more complex tasks and decisions in business and everyday life that still require human skills, expertise and creativity. AI can (only) help humans to make better decisions and to work in newer and smarter ways.
Here are 4 reasons why AI won’t take your job:
AI doesn’t have emotional intellect
Have you ever heard that soft skills, like communication and building relationships, are becoming more important for jobs nowadays than technical skills? This – emotions – is what makes us human. People expect you to share your emotions, be empathetic – for example in customer service. Sure, a robot can find an issue and suggest a solution, but would that interaction be enjoyable? Moreover, emotional commitment and a good relationship with the team members improves the employees’ engagement. Thanks to relationships we can find business partners and clients, as humans tend to work with people they like. This is the emotional side that machines can’t obtain.
AI can’t do creative work
Computers are great at repetitive tasks. Humans, on the other hand, really excel when it comes to creative thinking. Evidently, AI can appear to be creative, if you view creativity as just another skill that can be learned. AI needs to be trained, fed thousands of songs or poems and it will find some patterns and will try to recreate, but can we really call it creativity? Computers are currently not able to create truly random and original data.
AI can’t think outside the box
As mentioned at the beginning of this article, the narrow AIs available today are programmed to do only one certain task. If anything comes up that is not within what the machine was trained to do, the machine won’t be able to do it. AI can only process the data that it has been fed and that was designed for it. Humans still need to define the cases and scenarios that the AI program will operate under. An AI program will work within those cases and scenarios, but it will not define new scenarios to operate in.
Someone needs to program the AI
As simple as it is – someone needs to manage the AI, train it, set up the tasks, maintain it. The same thing happened during the industrial revolution, in the late 18th and early 19th century, when mass-production and assembly lines came into use, new job positions opened like maintenance and controlling. With AI coming more into our life, there will be new jobs like AI and Machine Learning Specialists, Digital Transformation Specialists, Software and Applications Developers, Robotics Specialists and Engineers and much more.
Why should you be excited about AI
We are still far away from developing General AI which could perform any intellectual task with efficiency like a human. Currently we have only narrow AIs that can be trained to do one narrowly defined task. An AI never gets tired and can keep working, processing useful data for you without breaks. Its calculations are error-proof and itis perfectly suited to perform lower-level routine tasks that are repetitive.
Recent Forbes article “AI 50 2021” presents a list of America’s 50 fastest growing and most exciting companies. Among the top 50 are startups that help doctors screen for cancer, help people learn languages, and help farmers get rid of weeds. Some track cargo ships as they sail the oceans, and one helps construction workers install drywall. These algorithms are going to make us safer, smarter, more efficient and more productive. But none of them threaten to replace doctors, teachers, farmers, sailors or construction workers.
At Iris.ai we have developed the tools to help researchers do literature reviews faster and more efficiently, so they can save time and focus on other tasks. Our tools find more relevant papers and can reduce manual work by up to 80%.
Many AI applications are helping humans to do their job better, to make our life easier. History has taught us that after the so feared Industrial Revolution we became more productive, the costs of production decreased and the quality of life improved. Therefore, the AI revolution will likely –we believe– upgrade our standard of living and open up more options.
Key Takeaways
AI can help us be more efficient and more productive. It will take over some jobs, but it will create many new ones. Humans beat –and will likely continue to beat– machines with their emotional capabilities, creativity and thinking out of the box.
Our brains think linearly, and we expect the future to be one step better than the past. Thirty linear steps would follow the sequence 1, 2, 3, 4 and so on, and we’re wired to think tomorrow’s technology will only be linearly better than today. However, digitized technology works exponentially. If you were to take 30 exponential paces, meaning 1, 2, 4, 8 and so on, you would literally circle around Earth 26 times. The future of R&D is exciting!
We cover large ground in this post, so feel free to skip to the section most relevant to you. We’ll talk about the various areas of R&D, in which we believe technology will have an impact: software as a strategy, predictions of chemical reactions, labs on demand, laboratory automation, supply chains, team structure and re- and upskilling.
Predicting the future of R&D
Software as a competitive advantage (or disadvantage)
Satya Nadella, CEO of Microsoft, says ‘Today no company can survive or thrive without a foundation, an infrastructure and a strategy built on software.’ It’s not enough to just find some software and apply it for processes, it has to be a fundamentally thought through system and strategy for that software. It is fundamental for successful R&D in companies. For example, in Tesla, cars can change physical behavior via a software update, which is radically changing the concept of a car and giving Tesla an advantage until all the others catch-up.
Building a layer of software as a foundational strategy for your company is what places you in the right position for the future. The cost of starting a software company has dropped significantly in the last few years. In the 90s it was $10 mln and today it is only $35 000. It means that if you don’t go out and take ownership of the software part of your business, then chances are someone else will. The digital foundation has to be an essential part of the strategy for every R&D team that wants to stay relevant for the future.
Predictions of chemical reaction before picking up a pipette
In the future, it will be possible to have predictions of chemical reactions before picking up a pipette. There are already papers that predict the outcome of chemical reactions with much higher accuracy than trained chemists. Thanks to major breakthroughs in sensor technologies we can actually start building digital twins of labs or physical assets and we can try things in the digital realm before implementing them in real life.
Laboratory automation
It is estimated that the market for laboratory automation systems will double from 5 billion to 10 billion by 2026. As we merge the field of AI and robotics, we will have smarter lab systems, saving labor time and costs.
Labs on demand
As more lab work is fully digital and labs being automated, you can end up not needing to be physically in the lab. It means that the lab could be absolutely anywhere. Any advanced and expensive lab equipment could be shared between multiple companies or research groups, which would drastically reduce costs of R&D. If doctors can already do long-distance heart surgery, there is no reason why a robotic lab couldn’t be shared remotely by a number of different researchers.
Your team will change
There are four specific areas where we believe that teams will radically change: fluid organizations, team autonomy, crowdsourcing and startups, and agile players. All organizations have to become more fluid. Employees must be able to work with more partners and not as one rigid hierarchical organization anymore. The R&D team will have more autonomy to collaborate with different entities, and the future teams will not be about hierarchies but about networks.
Continuous re- and upskilling as requirement
Constant reskilling and upskilling will define a successful company. The famous researcher, John Seely Brown, said that the half-life of a learned skill is five years. Therefore, employees will have to learn how to adapt to new tasks to better perform. With new technologies emerging rapidly, companies must be able to adapt to stay on top. Every next step will require a different skill set. Even if the process might still be the same, the skills needed to navigate certain cases radically change.
All the insights at your fingertips
Many people realize that there is too much knowledge and information and we need better ways of handling it. The idea is to have an AI that we can have a conversation with to help you find exactly the information that you’re looking for. This is at a very early stage now, but it is something that is developing rapidly.
The research papers and patents of the future will look nothing like today
Blockchain is one potential area in this realm where it’s a system that can track every contribution made to science, or anything that was written. This system can track without changing ownership of an idea. Then it could distribute the financial outcome of those ideas accordingly. That could radically change the pace of innovation.
Key takeaways
The future is already here and if we don’t continuously embrace it, someone else will jump in and build something better. We are a part of this future and there are going to be interesting times ahead.
In pharmacovigilance, post market surveillance and many other areas of drug development, highly skilled team members spend an enormous amount of time systematically mapping out publications. This includes going through clinical trial reports and real world evidence, and extracting key data points such as adverse effects.
In this webinar, we covered how to:
👉 Automatically filter based on entities of interest and specific context.
👉 Extract data such as adverse effects, treatment and patient baseline characteristics.
Artificial intelligence is rapidly changing the way we work in various companies and industries around the world, including the chemical industry. Organizations are adopting these technologies to accelerate processes and reduce costs, as well as saving employees from tedious, mundane tasks.
Accenture suggests that there are three ways of applying artificial intelligence in research across industries:
Reinventing the process to manage process change, rethinking standardized processes as continuously adaptive, and using AI across multiple processes.
Rethinking human-machine collaboration; how companies can have an AI-enabled culture to reskill employees to work in alliance with machines.
Utilizing data, making use of AI and data to solve previously unsolved problems and reveal hidden patterns.
In this article, we will explain how chemical researchers are applying artificial intelligence.
How chemical researchers are applying AI
There are three categories of chemical research that are affected by AI. The first category is molecule prediction — draw on known properties to predict new behavior. The second category is synthesis models, which predict how to create certain molecules in fewer steps and more reliable processes. The third is handling prior knowledge to make sense of what we already know —starting with data mining to find the right information.
1. Case studies on molecule predictions
The pharmaceutical industry is one of the front runners in AI. In February 2020 the model in “A Deep Learning Approach to Antibiotic Discovery” was created, a model that translates molecules into vectors. It starts with every atom being represented with a vector of simple properties. This is used to create a fingerprint of the molecule’s structure, which helps the neural network to learn.
The model was trained on tests with E.coli to see what molecular structures actually were antibiotic. Then it was applied to the Broad Institute’s drug repurposing hub – an open-access library of more than 6000 molecules with known biological activity. As a result, they discovered a compound called Halicin with impressive antibiotic activity, despite having a chemical structure unlike conventional antibiotics.
Following this success, the team applied their AI technique to a database known as ZINC15 — 107 million molecules were manually selected for screening. Based on the deep learning tool’s predictions, 23 compounds were chosen for further investigation. Two of these compounds showed promise against a range of drug-resistant E. coli.
In march 2020 Münster University published A Structure-Based Platform for Predicting Chemical Reactivity. The new tool is based on the assumption that reactivity can be directly derived from a molecule’s structure. It uses an input based on multiple fingerprint features as an overall molecular representation. Organic compounds can be represented as graphs on which simple structural (yes/no) queries can be carried out. Fingerprints are numeric sequences based on a combination of multiple queries. They were developed to search for structural similarities and proved well suited for use in computational models. For the most accurate presentation of the molecular structure of each compound, a large number of different fingerprints are used.
2. Finding the best synthesis method: expert system vs. machine learning
In 2018, The Defense Advanced Research Projects Agency (Darpa), the development agency of the United States Department of Defense, presented a project where artificial intelligence was used to develop and find the best synthesis methods. The user can input any structure, either known or novel, and then a machine generates thousands or even millions of reaction sequences in order to end up with the final product. Reactions are being ranked and identified based on feasibility, cost, and other factors. Darpa has two ways of doing this. They can apply the expert system, a system based on 60 000 handwritten rules, which is effective but not scalable. Alternatively, they can encode each of the molecules to predict bond changes, using machine learning (much like on molecule predictions). The next step is having a manual help to filter the results and generate a shorter list of top candidates.
There are three fundamental problems in using the machine learning approach as opposed to the expert system. First, the challenges seen in using machine learning, in this case, is the data acquisition. There is missing information and biased reporting due to lacking reports on failed experiments. When it comes to reaction sequences that can be extracted from patents, not all information is going to be reported in the same place.
The second disadvantage is data representation, meaning how this data is presented and explained to a machine in a comprehensive way. The data format needs to be considered and determined — whether the data is presented in formulas, images, features, properties, etc.
The third problem is the exploration space. That space is so much vaster than the information we have available. That raises questions about how to teach a chemistry engine to invent new potential molecules and pathways when we don’t have data on that at all.
There is a model called Molecular Transformer: A Model for Uncertainty-Calibrated Chemical Reaction Prediction which can predict the outcome of a chemical reaction with much higher accuracy than trained chemists, and it will suggest ways to make complex molecules. However, it needs a lot of data and in a very specific text-based format called SMILES (simplified molecular-input line-entry system) that has been data mined from patents. In the end, the preparations to use it for a specific use case might not be worth it from a cost perspective.
3. Organizing knowledge
Artificial intelligence is already used in prior art. There are a few existing and future inventions in that area which will change the current process radically. The first and most basic invention already in use is smarter search. Automated literature reviews is the second step, which we have been working on for the past five years at Iris.ai. We’ve gotten to semi-automation, meaning the search needs human-machine collaboration.
The next frontier that we are working on is identifying specific insights from text. The first step is advanced data extraction and linking, which we have developed in our Extract tool. The PDF to be extracted is sent to the Iris.ai system. This PDF can be a patent, a clinical trial report, a research paper or any other relevant type of scientific content. It can be one simple document at a time, or hundreds or thousands of them in a batch. The Iris.ai engine extracts the text and identifies all the domain specific entities, then locates the tables and extracts the data from rows and columns, and links the data between the text and table. Graphs, figures and other elements go through the same process. Then the engine populates a pre-defined output in a machine readable format; an excel sheet, an integrated lab tool, a database or anywhere else your researchers require.
What’s important in this step is the self-assessment module which communicates to the human researchers how confident the machine is in its results, to give the human guidance on where to do the most rigorous manual verifications.
In the long-run, we expect to see developments in hypothesis extraction from the prior art, knowledge validation based on prior art, and lastly, drawing new conclusions and finding new hypotheses from all of the existing prior arts.
Automating manual tasks vs. rethinking the imaginable
There are two very different mindsets when it comes to applying AI in your organization. You can replace a human process and have a machine do the same activity but faster, for example, in extracting the data. Willingness to invest time and resources is needed, but there is clear ROI and known outcome and benefits. The second mindset is about activities that cannot be done by a human. For example, a machine can identify new potential application areas, meaning you need willingness to invest as well as rethink and re-imagine what’s possible (ROI will be unknown until you try).
Interpretability and explainability
One of the emerging fields in AI worth mentioning is interpretability or explainability. It is not just AI that tells if something will work or not, but explaining why. For example in molecule prediction, AI can predict that certain actions will cause an activity or property because of a specific area in the molecule or combination. As a result, it gives the chemist an immediate indication of how it could be altered if the reaction is unwanted. Similar to the data extraction tool that Iris.ai is working on, where every row and column will come with a machine-created self-assessment with a percentage of certainty.
At Iris.ai we have spent the last five years researching and developing an AI Engine for Scientific text understanding. Already successfully deployed in a generalized suite of tools for Academic literature reviews, we believed it was time to see how this engine could be reinforced on one specific domain, and how it could be used to find precise and more spot-on answers for industry researchers. Chemistry was an interesting place to start, for the reasons outlined below, as well as because it is an industry ripe for digital innovation and essential for the sustainable future of our planet.
The interesting thing about chemistry
In 1776, chemist and mechanical engineer James Watt invented the Watt steam engine, which was fundamental to the changes brought by the Industrial Revolution. Ever since – and potentially even before – an understanding of chemistry has been the foundation for our technological development, and there is no reason to believe that this holds any less true for the future. Whether we need more sustainable materials or biodegradable fuel to reduce our carbon emissions, new materials allowing us to travel to space or terraform Mars, novel ways of ensuring that every person on this planet is properly fed or understanding how we can handle an ocean filled with plastic particles, chemistry is going to be absolutely foundational.
What has enabled such a thorough understanding of chemistry pertains to the field’s formalism – the same as for maths and physics. This means structured approaches to unifying language so that any chemist anywhere can talk about anything from the basic elements, via molecular formulas to complex synthesis procedures in the same way. This structured way of communicating with each other has allowed rapid progress in this scientific field.
However, formalism has its downsides: when you simplify a process or a thought process into a unified language, inevitably there will be a loss of information on the way. Much like a compressed image is easier to share and still show the same motive – but is pixelated, so can formalist research results be easier to convey transmitting a general idea of the approach, whilst missing the finer details though. Ideas are compressed to formulas, long research papers compressed to abstracts, novel ideas compressed to a 140 character tweet, detailed lab notes compressed to summaries.
In chemical research, this ‘compression’ has been required because of human limitation – but today, it isn’t required anymore. Computers have already allowed a much broader and larger volume of shared knowledge – which in itself makes absolute formalism tricky. And thanks to advances in AI, we are rapidly approaching a new frontier in chemical research (and beyond).
With new AI advances, machines can help researchers find what other researchers have done, ‘translate’ it into that researchers’ current context, and get a much higher clarity on how and why the solutions or conclusions were reached – without the information loss built into the current process. The machine will have all necessary information as there is no information loss – but only communicate or ‘translate’ the exact relevant pieces between the researchers. This will truly be a new paradigm of chemical research, and we intend to be part of it.
Iris.ai’s first steps into chemistry
We have taken our core engine and reinforced it on chemistry. The interesting thing about that approach is that because the starting point of the general engine is strong, we only need a small collection of research paper in the specialized field, or for some use cases a seed ontology already created by a human, to specialize the tool – which makes it very flexible and re-deployable on many different research fields with similar user needs.
We are already now building out this engine into the first set of tools that will help Chemistry researchers on three different levels:
Discover. When dealing with unknown unknowns, the Discover tool allows interdisciplinary discovery, beyond today’s limiting keyword queries. It fingerprints the description of the researchers’ problem, and maps out all relevant papers and patents they should be reading to get a full overview of the field. The discover tool is especially helpful at the early phase of a new and interdisciplinary project, where it has proven to give researchers a better overview, find more spot on papers and draw better conclusions.
Identify. When the researcher knows the answer is ‘out there somewhere’ but it’s like looking for a needle in a haystack. Known unknowns can be found through this conversational AI that guides the researcher through the information found in millions of documents, asking the right questions to narrow down to exactly the bits of knowledge you need. This knowledge could be finding new application areas for an existing compound, identifying better synthesis procedures or simply identifying the right material for your use case.
Extract. In spite of chemistry being such a formalist field, every researcher writes in their own way. That means when you have a need to extract key data from a document – for example experiment data before going to the lab to recreate them – it takes a lot of valuable researcher time. Our automatic extraction achieves 90% accuracy and perform two months worth of manual labor in a matter of minutes.
At Iris.ai, we are very excited in bringing our AI skills together with some very talented chemical researchers, and see what just might be possible when you bypass the limitations of human formalist language and let an AI understand the context of your words to help advance your research.
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.
Iris.ai puts her engine to the task of COVID-19
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.
