AutoML can solve many of the problems businesses face, but not all. Democratization of machine learning can make a difference. Adam Carrigan, co-founder and COO of MindsDB, explains.

Machine learning and predictive analytics present a massive opportunity for businesses looking to improve existing processes, become more data-driven, and enhance the customer experience. Many companies have instituted some ML projects and are ready to take the next step with more widespread adoption.

Unfortunately, this is one area where companies run into challenges. The expertise and investment required to make ML and predictive analytics successful aren’t readily available. The demand for data scientists continues to snowball. According to LinkedIn, the data science field saw 37 percent job growth last year.

How can companies deal with this shortage of data science resources to take full advantage of AI and predictive analytics? We talk with Adam Carrigan, co-founder and COO of MindsDB, to learn about adding automation to machine learning.

Upside: What are the current challenges with adopting machine learning?

Adam Carrigan: Now that the use of machine learning and predictive analytics is more widespread and investment in this technology is increasing, companies are finding it difficult to scale the benefits to the entire organization. The department that oversees machine learning (often the data science team) has, in many companies, now become a bottleneck within the process to implement these solutions.

Data science teams are usually small, and the process of researching, testing, and deploying machine learning can take anywhere from days to months depending on the complexity of the data. This results in a long list of backlogged projects, where specific projects take priority and others are pushed back or never implemented. This challenge is even more significant for SMBs, which typically don’t have the resources to invest in highly in-demand data scientists.

AutoML now attempts to automate specific tasks in the ML workflow to reduce the strain on data scientists, but this comes with limitations. Many of the solutions available today still require a data scientist to play a significant role in the process. Although many aim to automate the training and testing, this leaves much of the process untouched, including the crucial deployment stage.

How should companies approach automating machine learning?

There are two approaches to automating machine learning inside an organization. The first is automating more of the process for the data scientist, freeing up capacity for them to solve more problems faster. This increases capacity and at least temporarily solves the bottlenecks I talked about earlier. This approach doesn’t help those who don’t yet have a data science team and is primarily a Band-Aid solution.

The second approach, and the one I believe will have a more meaningful impact, is to equip existing employees with the capabilities to solve most machine-learning problems. This eliminates the need to go back and forth between teams because ultimately the team trying to solve the problem is the one that knows it best. This is great for companies without a data science team to begin with. This approach also has an added benefit for those companies already with a data science team. It frees up their time and resources for the more complex and challenging problems. Ultimately, AutoML can solve many of the problems businesses face, but not all. In many instances, data scientists are still essential.

This approach requires a fundamental shift in the way we approach machine learning. Instead of thinking about it as an element within the application layer, it becomes useful to think about it as a data-layer problem. This opens up some exciting opportunities.

When you consider ML as simply other representations of your data, it makes it easier to run models and predictions at the data layer. AI tables — automated ML models as native data tables inside databases — let users execute models by merely running a data query. Instead of using AutoML to streamline the data scientist’s tasks, this approach puts ML into the hands of the end-user of the data.

How can companies educate the end users of predictive analytics?

This is a crucial component to the democratization of machine learning but shouldn’t necessarily be treated any differently than any other method of improving the skills of the enterprise’s employees. These users still need basic education to use those tools. Essential skills include running queries, analyzing data, and understanding how to use ML data to support the human decision-making process.

Now that machine learning is simple enough for the average database user to leverage, these users must now understand the potential strengths and weaknesses in the models they produce. Teaching end users how to determine a level of trust and confidence in the models using explainable AI (XAI) is even more important than the operational education. Thankfully, more AutoML solutions are including explainability as a standard feature. Armed with this information and the generated predictions, users become even more powerful in making impactful changes within the organization.

The post Advice for Automating Machine Learning and Predictive Analytics appeared first on Artificial Intelligence.

Organizations should consider adopting AutoML to ease the process of data analytics by automating the process.

Industries have been leveraging AutoML to enhance data processing and data engineering. However, there are discussions of how AutoML will affect the job of data scientists. Let us understand more about this technology and its role in enhancing the data efficiency of a company.

The digitization and automation across organizations demanded the adoption of data science and advanced data analytics to encourage business growth and agility. With this increased pace of transformation, companies started to employ data scientist teams to address the need for developing machine learning models and analytics algorithms.

Data-driven decision-making in organizations has proved to improve productivity and minimize costs in the long run. Due to the highly technical skills required for the job, the supply of data scientists is limited even now, thus making it difficult for organizations to capitalize on data and create machine learning models to analyze them. This is where AutoML comes in.

Why AutoML?

Automated Machine Learning is a nascent development in the field of artificial intelligence. AutoML automates the end-to-end machine learning requirements in business operations. This technology enables the development and deployment of machine learning models without any time or skill constraints.

The conventional procedure by data scientists takes a good portion of time since it involves data cleaning, data analysis, identifying machine learning models, running them, conducting parameter tuning, designing the algorithms, and deploying them. Integrating this long process into the workflow of organizations can be difficult and time-consuming. Since there is a shorter supply and high demand for data scientists, it becomes tougher to develop a team.

Automated Machine Learning eliminates all these challenges by automating the process and running several machine learning models at the same time. AutoML also aids the process of feature selection, feature extraction, and feature engineering to run algorithms. The amount of data is increasing each day and so is the adoption of big data in organizations. Hence, AutoML is a desirable technology to reduce the time and complexity in the implementation of machine learning models.

Another commendable benefit of employing AutoML is its role in the democratization of data science in organizations. There is a huge skill gap in most companies concerning the high skill demand for data science. Organizations usually find it difficult to address the need for better machine learning models because of the limited access of people to the field of data science. AutoML for organizations eliminates this gap by encouraging ‘citizen data scientists ’ to perform the tasks without any prior expertise.

It enables employees other than people with data scientist qualifications to contribute to the data science ecosystem with minimal assistance from the data science teams. For example, Cloud AutoML by Google enables businesses to build customized machine learning models with limited skills and expertise in the field. AutoML increases the accessibility of data science and data engineering to a larger audience rather than restricting it to a popular group.

Will AutoML Eliminate Data Scientists?

If you want a single-word answer then, No-AutoML will not make data scientists disappear. It will ease the burden on the shoulders of these data experts by taking over repetitive tasks that do not need much attention. AutoML will automate some of their tasks and leave them with those that need highly technical skills. Organizations will still need data scientists to define problems, apply domain knowledge on the issue, and generate reasonable and creative models. AutoML can work alongside data scientists to support them and this course will enable the decentralization of data science knowledge.

The post AUTOML: AUTOMATING AND DEMOCRATIZING DATA SCIENCE IN ORGANIZATIONS appeared first on Artificial Intelligence.

There is an enormous amount of data available for companies for business insights and analysis. Businesses now have tools to enable them to aggregate them into a meaningful report in order to use it for decision making. Combining big data in a meaningful way is tricky – but bigger brands are tackling it well.

A majority of organizations are aiming for a data-driven approach, and are seeing success in their efforts. According to a study by Dell EMC (back in 2014), there would be 1.7 megabytes of data produced in 2020 – for every person and in every second.

Businesses need to follow proven practices for data aggregation to reduce related data management challenges. A major issue is ensuring that they are running on the raw insights, and organizations can accomplish it by structuring and normalizing data. Thus, to begin with, the top methodologies need to be executed.

Basically, businesses need to realize their short-term and long-term analytics objectives. For instance, currently, a company could be trying to know its consumers’ buying preferences. After a while, it may want to aggregate data from different sources to identify audiences’ interests – in order to sell insightfully. Regardless of the purposes, there is likely to be an immediate and long-term focus that will alter the business’ data aggregation requirements. And the strategy should reflect it.

For organizations that purchase data from third parties, they need to ensure that their privacy standards and governance are compatible. In this case, healthcare data would be a great example. While acquiring patient data from an external source for sensitive issues for the purpose of analysis or treatment, the data needs to be in an anonymous format. This is to secure the privacy of such patients.

Furthermore, businesses need to determine how data will be accumulated and how the users will be accessing it. The aggregated data can be used by specific functional areas in a company or by different departments across the board. This is a critical factor because it indicates the best choice- whether the company has chosen to aggregate and keep data in a vast data repository with various access choices – or in a small database that is customized to the need of a specific user group.

In its essence, automating data integration will help. No matter where the data is being aggregated, organizations will require a straightforward way – to vet and integrate the data into the target data source. The necessity of having to hand-code the data integration interface needs to be avoided. Hence, the preferred tactics for data integration are generally processed via standard APIs and automated integration solutions tools – in order to perform secure data integration for business functionalities.

The post Big Data – How Businesses Can Manage Data Aggregation Successfully appeared first on Artificial Intelligence.

Robots and robotics, the two ambiguous terms, have gained more traction after the onset of the COVID-19 crisis. From automating the industry chains to monitoring the public crowd, the resourcefulness in the pandemic is undeniable. Before that, robots have already worked in environments or job lines that are deemed hostile or beyond the basic scope of humans. In short, they have always been there in our hard times and have enabled us to move towards a digitally enhanced future.

During this global health emergency, robots have helped us either by carrying health samples, distributing ration, test kits, sanitizing public spaces, conducting surveillance, and many more lifesaving acts. Denmark-based Blue Ocean Robotics has developed an autonomous mobile robot that can enter a room and disinfect it with UV-C light, without exposing medical co-workers to the potentially harmful radiations. ABB’s dual-arm mobile YuMi robot can move its way around people using attached sensors. Its functions include deploying medicines, bed linens to patients, etc.

The adoption of robotics in the pharmaceuticals and drug discovery is expected to grow from US$2.05 billion in 2019 to US$4.87 billion in 2023, growing at a CAGR of 18.9 percent. Now, robots are ready to help the healthcare industry by assisting in drug discovery too. Although robotics and robots in pharmaceuticals is not a new topic, the coronavirus pandemic has pushed for the mainstream usage of the same in this industry.

Berkeley University is using robotics to leverage its pop-up testing lab. An automated liquid-handling robot analyzes swabs from patients and to diagnose them for COVID-19. Andrew+ a pipetting robot, from Andrew Alliance S.A., a Geneva, Switzerland-based company, can help to carry out repetitive manual experiments and tasks. All these robots have been helpful to humankind to fight against COVID-19 while allowing them to maintain social distancing. In July, Insilico Medicine, a biotechnology company developing an end-to-end drug discovery pipeline utilizing next-generation artificial intelligence, had announced a partnership with Arctoris, the world’s first fully automated drug discovery platform. The objective of this partnership is to discover, synthesize, and profile a set of inhibitors for COVID-19 treatment.

The process of discovering any new drug that can cure a disease is like finding a needle in a haystack. Further, it is an incredibly expensive, arduous, and time-consuming solution. The conventional pipeline for discovering new drugs can take between five and ten years from the concept stage to being released to the market, costing billions in the process. The pharmaceutical industry is increasingly making use of robotics to automate specific processes into drug development. This includes drug screening, anti-counterfeiting, and manufacturing tasks. Along with Artificial Intelligence and machine learning, it can help ramp up the speed of drug discovery.

Besides, most of the tests performed in laboratories regarding research and development of drugs typically require conducting rote tasks like moving fluids and transferring them into test tubes. Swiss multinational pharmaceutical company Novartis notes that earlier running samples by hand had a potential output of around 30-40 per day, where after upgrading to automated robotic systems, the output has increased to hundreds of thousands per day. Today, processes such as nuclear magnetic resonance (NMR), mass spectrometry, high-throughput screening (HTS), cell culturing, and high-performance liquid chromatography (HPLC) can be carried out by robotic arms. Meanwhile, for liquid dispensing, Sygnature uses the Labcyte Echo acoustic dispensers to prepare assay test plates using nano-liter volumes of compound solutions. In this way, robots make an ideal choice for such jobs since they are easy to automate and provide a high level of accuracy and consistency.

The biggest advantage of harnessing this technology is that robotics minimizes the risk of human contamination. Due to their ability to accommodate a wide range of products that require constant modifications, robots are used as integrators and tool changers in configured reactions. Moreover, they are employed (e.g., barcode scanning bots) to identify counterfeit drugs and medication. Apart from that, with the introduction of robots in laboratory automation system sizes can be reduced as guarding is eliminated.

So, one can observe how robotics and robots have impacted the drug discovery, development process in the healthcare industry. These technical applications of modern disruptive technologies have greatly enhanced the ability to explore and identify new drug candidates. And, COVID-19 gave the pharmaceutical sector the much-needed push. Soon scientists, hunched over Petri dishes, trying to extract new compounds with possible medical uses, and conducting research and analysis by hand will become a memory of previous decades. A new era is awaited to bring a monumental change.

The post HOW ROBOTS AND ROBOTICS ARE POWERING THE DRUG DISCOVERY PROCESSES? appeared first on Artificial Intelligence.

The Legal Ombudsman (LeO) is exploring whether to use big data and machine learning technology to suggest outcomes and make recommendations in resolving complaints about lawyers.

The plans were revealed in a collection of articles produced by the Legal Services Board to solicit views from regulators, regulated groups and others about how regulation could support innovations in technology to improve access to justice.

Senior ombudsman Mariette Hughes, who is is charged with improving the watchdog’s efficiency, said a first step might be to use technology to triage decisions about the risk and complexity of incoming cases in order to know which staff resources to devote to investigations.

“Automating this area of work would mean things move more quickly through our process and would free up staff to focus on the investigation of complaints,” she said.

In future, machine learning technology could be used “to identify patterns and trends between complaints made and outcomes reached and support [LeO’s] objective to share learning and insight from cases with the profession”.

She continued: “The technology could be used to suggest outcomes and make recommendations for resolution, based on the trends and correlations identified in existing data sets.”

However, she cautioned that maintaining the human element in investigation was vital: “It is important to remember that effectiveness is not always about speed; it is also about accuracy and freeing people up to undertake the complex elements of our work, and to have the time to build relationships that can ultimately support the resolution of complaints.”

Meanwhile, LeO has issued guidance on its approach when considering complaints as a result of the difficulties brought about by the coronavirus, making clear that it would be flexible.

It recognised that some lawyers were operating “a reduced service”, while others have closed altogether, while “many complainants do not have the capacity to follow a complaints process at the moment”.

While LeO still expected lawyers still to complete investigations into any complaints within the eight weeks they are normally given, being open and honest with clients about delays, and keeping evidence of the problems faced, would be taken into account.

It was the same for those struggling to respond to a LeO investigation, to the point where one option would be to suspend it until the situation improved.

Lawyers would also not be penalised for refusing to take on a client if they were complying with government guidance, or their business has been “affected with levels of absence which mean you are unable to accept new business”.

LeO said it would expect the six-month period for clients to make complaints to be sufficient for most people, even under current conditions, but that it had the discretion to extend this if that was deemed appropriate.

The post LeO weighs using big data to aid decision-making appeared first on Artificial Intelligence.

At first, you might find the idea of Artificial Intelligence/Machine Learning being associated with Law very unlikely since both the fields appear to be poles apart. Well, the truth is far from it; today, Artificial Intelligence or AI is on the way to transforming the legal profession in various ways, helping Law firms manage their operations as well as augmenting and reducing many of the tasks that were previously relied upon humans to do, saving precious time and manpower that can be otherwise used for more productive tasks.

Artificial intelligence, on the base level, aims to develop and find ways to reduce, manage and execute laborious tasks for different industries, automating many of the operations that would otherwise require human input. These solutions are found to greatly improve the speed as well as drastically reduce the errors, hence improving the accuracy.

Whenever a new technology is introduced to the World, every sector and industry are offered the possibility to adopt that to enhance their operations. One example is of computers and how they quickly grew in use, taking over many of the manual paperwork, and how they have become essential today in almost every office and profession. Law firms are no exception to that, where technology has always been forefront and finds its way into supporting the lawyers, paralegals, researchers and clients alike which are associated with the profession.

Through this article, we will be looking at and discussing the different ways in which AI and machine learning are applied specifically in the legal profession for the streamlining of their operations and work processes. Let us start by briefly understanding what artificial intelligence and machine learning truly are.

What is AI and Machine Learning?

The term Artificial Intelligence is meant to describe intelligence exhibited by machines that try to copy or mimic the human cognitive in certain ways, such as problem-solving and learning. This concept is widely being accepted and utilized to substitute those tasks that require human intelligence.

Machine learning denotes the phenomenon where computers are fed with data sets and are made to analyze different patterns from that data, learn from there and then use that learning to gain useful insights. Different algorithms are involved which aim to let the computer carry on the task without the need for human intervention.

The Impact of AI on the Legal Profession?

There is no doubt that AI has already started impacting the legal sphere in ways you might have noticed as well (if you belong to the profession, that is). Research has shown that a lot of manual work required in Law firms have been substituted by artificially intelligent machines. Since the operations are greatly improved and streamlined, as per different studies, Law firms have no choice but to embrace this new wave of technology and accept AI and machine learning in their daily operations. Below, we describe some of the different applications that Law firms have found with AI.

Documents review 

A great part of any legal proceeding is the amount of documentation that needs to be skimmed through to find the relevant material, which requires a team of paralegals and hours and hours of precious time. Firms have adopted AI software that helps to analyze documents and flag the ones that are deemed as relevant.

Legal Research 

Once the relevant documents are shortlisted and flagged, machine learning comes into work and uses the learned algorithm to find similar documents that can be of use, out of the millions of papers, proceedings, and dissents. This ensures that all the legal research is done very efficiently and in a comprehensive way, covering a whole lot more data. Natural language processing or NLP is also applied to further analyze the documents to aid the research.

Due Diligence 

This is hectic work, where legal professionals are required to perform exhaustive background checks and go to lengths to uncover information regarding their clients or on their behalf. These facts and data are then evaluated for better decision making and are used to support their cases and giving sound counsel to the clients. This tedious work is fast being replaced by artificially intelligent systems, which help perform most of the due diligence more efficiently and accurately.

Contract Management 

Often, clients come for legal counseling to review contracts and to check and identify if any issue or risk is associated with that contract. Sometimes, contracts can be misleading and have a negative impact, and legal professionals assist their clients to avoid just that. AI can be used and is being used, to analyze such contracts in bulk and made to learn to identify such situations quickly with fewer human errors, avoiding any mishap. 

Predicting Outcomes

Since computers and artificially intelligent systems have access to huge amounts of trial data and years of documentation, pattern recognition, and machine learning comes into play in the analysis of all that. These insights are then used to predict outcomes for similar cases as well as finding answers that could have otherwise taken days to arrive.

Conclusion

According to recent research, many and more of the legal roles will be automated and be replaced by artificially intelligent systems. The time is ripe, therefore, for Law firms to accept and embrace this new technology change or watch as others who do fly past.

Already, Artificial Intelligence and machine learning, hand in hand, are transforming the legal sector where organizations have come to realize that technology and innovation is the key to success. Already, these intelligent systems have showcased their superiority over humans, statistically speaking; numerous reports and studies will tell you just that. Especially for Contract management and legal research have shown a lot of promise.

Even though some would argue that manually reviewing documents can prove to be accurate and humans can compete with the threshold that is left by the machines, there is no doubt about the fact that AI systems are proven to be no match when it comes to speed. They get the work done in hours which would have otherwise taken a team of paralegals to perform in days!

The post HOW AI AND MACHINE LEARNING ARE TRANSFORMING LAW FIRMS appeared first on Artificial Intelligence.

Google LLC today debuted some important updates to its Dialogflow, the main technology that powers its Contact Center AI service for automating interactions with customers in call centers.

Dialogflow is a conversational artificial intelligence engine used to create virtual agents that can understand and respond to all manner of queries from callers, using both voice and text as a medium.

The main update today is a new “Dialogflow Mega Agent” in beta test mode that increases the number of “intents” available to virtual agents by up to 10 times, to 20,000 in total.

“Increasing the number of intents means more training phrases, actions, parameters, and responses to help your Virtual Agent interact with customers and get their issues resolved more efficiently,” Levent Besik, director of product management at Google Cloud, and Shantanu Misra, a Google Cloud product manager, wrote in a blog post.

Regular Dialogflow agents are limited to just 2,000 intents, but increasing the limit enables Contact Center AI agents to have more “natural and seamless conversations” and to pivot questions and intent when they want. As a result, Google said, customers can tackle more use cases in order to better serve their customers.

Other updates include a new Dialogflow Agent Validation tool that developers can use to identify flaws within the agents they create.

“It does this by highlighting quality issues in the Dialogflow agent design, such as overlapping training phrases, wrong entity annotations, and other issues, and giving developers real-time updates on issues that can be corrected,” Besik and Misra said. “Reducing errors leads to faster bot deployment, and ultimately, higher-quality Dialogflow agents in production.”

Dialogflow Versions and Environments meanwhile enables developers to create multiple versions of the same agent that can be used for different environments, including testing, development, staging, production and so on.

In addition, Google announced a new Dialogflow Webhook Management application programming interface.

“With Webhook Management API, you can now create and manage webhooks, making it easier for enterprises to programatically fulfill their queries,” Besik and Misra said. “As Dialogflow processes and fulfills millions of queries daily with webhook, this new API, which was previously limited to the Dialogflow console, will help enterprises speed up their agent design process.”

Google said the new features in Contact Center AI can be accessed today via the Dialogflow console or API.

The post Google updates Dialogflow AI engine to help customers create better virtual agents appeared first on Artificial Intelligence.

As ecommerce grows, so does the data that is stored and used. Only a fraction of that data is utilized, however. But that will likely change, as data scientists are getting better at merging, standardizing, and analyzing.

All of this will impact ecommerce merchants. What follows are my data predictions for 2020.

Impact of Data in 2020

Personalized stores. Merging search and purchase history of customers and lookalike visitors will create a much more personalized shopping experience. This will translate to higher conversion rates and more cross-sell opportunities.

Personalized marketing. Marketing will become increasingly sophisticated. Merchants will send multiple email variations based on customer segments. For example, if a customer buys only t-shirts, sending him an offer for pants will likely be ineffective. Similarly, customers who buy only discounted goods will presumably not respond to a full-priced offer. Marketing to both customer types requires collecting and segmenting the data.

Increased automation. Automating repetitive tasks not only saves human resources. It also improves the customer experience. An example is using chatbots for customer service, which can improve accuracy and response time.  In 2020, find ways to automate by asking each employee to describe repeated tasks. Keep in mind, however, that not all such tasks are candidates. Many have variations that require human intervention.

More cross-border sales. Automated language and currency translation, streamlined shipping (including customs), and local payment options will help merchants penetrate global markets with little investment. Even human translators (such as on Fiver) are becoming less expensive.  And shipping platforms and plugins can calculate at checkout the exact worldwide transit cost.

Better forecasting. Business intelligence tools can now forecast sales, optimize prices, and predict demand — in detail. The result is lower inventory quantities and targeted promotions based on a product’s demand. Businesses can move faster without spending a lot of money. To start, merchants can acquire an intelligence platform or hire a machine learning expert who can forecast in R or Python.

Research with social media. Marketers will focus on understanding the customer and her behavior leveraging the massive, public data on social media sites. Retailers will shift from using net promoter scores and surveys to analyzing qualitative and quantitative info. Merchants can start by manually categorizing the opinions of customers and prospects around products, product types, and the business overall. Over time this data can be aggregated for ongoing insights.

More privacy laws. Governments worldwide are imposing strict privacy laws on the collection and use of consumer data. Examples include Europe, Korea, and California. More will undoubtedly come. Merchants will spend money on legal fees, employees (such as data compliance officers), and consultants. Marketing capabilities will presumably decrease, as will customer experiences.

The post 7 Ways Big Data Will Impact Ecommerce in 2020 appeared first on Artificial Intelligence.

Apple has struggled in recent years to establish a robust artificial intelligence (A.I.) practice. This partially stems from the company’s ironclad privacy policies—it’s more difficult to analyze datasets for insights when internal rules prevent the company from using every piece of user data it can vacuum up. Nonetheless, Apple’s newest projects show that it’s powering ahead anyway—including one platform that, if it’s ever released, could change how you use A.I. and machine learning (ML).

(It’s worth remembering how, in a 2015 speech, Apple CEO Tim Cook accused tech giants such as Facebook and Google of “gobbling up everything they can learn about you and trying to monetize it,” which he framed as “wrong.” It seems unlikely that Apple’s stance on data and privacy will change during Cook’s tenure.)

According to a just-released paper with the dry-but-mysteriously-compelling title “Overton: A Data System for Monitoring and Improving Machine Learned Products,” a group of Apple researchers describe their work on a machine-learning platform (named—you guessed it—“Overton”) designed to “support engineers in building, monitoring, and improving production machine learning systems.”

How does Overton go about this herculean task? By automating the nitty-gritty of machine-learning model construction, deployment, and monitoring. Apple claims that the platform is already in use, supporting multiple initiatives “in both near-real-time applications and back-of-house processing.” These Overton-powered applications have “answered billions of queries in multiple languages and processed trillions of records reducing errors 1.7 – 2.9x versus production systems.”

This means that any researcher or engineer working with Overton will need to trust that the platform can recognize and fix issues with a model; otherwise they’ll presumably need to dig into the algorithms and datasets themselves, a lengthy and stressful process. But if it truly works as it says on the proverbial tin, it should reduce the time necessary to churn out results. Here’s an except from the paper on what the model inputs:

Overton takes as input a schema whose design goal is to support rich applications from modeling to automatic deployment. In more detail, the schema has two elements: (1) data payloads similar to a relational schema, which describe the input data, and (2) model tasks, which describe the tasks that need to be accomplished. The schema defines the input, output, and coarse-grained data flow of a deep learning model. Informally, the schema defines what the model computes but not how the model computes it: Overton does not prescribe architectural details of the underlying model (e.g., Overton is free to embed sentences using an LSTM or a Transformer) or hyperparameters, like hidden state size.

And this is what Overton does with that schema/input:

Given a schema and a data file, Overton is responsible to instantiate and train a model, combine supervision, select the model’s hyperparameters, and produce a production-ready binary. Overton compiles the schema into a (parameterized) TensorFlow or PyTorch program, and performs an architecture and hyperparameter search. A benefit of this compilation approach is that Overton can use standard toolkits to monitor training (TensorBoard equivalents) and to meet service-level agreements (Profilers). The models and metadata are written to an S3-like data store that is accessible from the production infrastructure. This has enabled model retraining and deployment to be nearly automatic, allowing teams to ship products more quickly.

So Overton is going to reduce the amount of coding that machine-learning researchers and data scientists need to do—allowing them to observe and manage the process from a higher level. Plus, it’s interoperable with platforms such as Google’s TensorFlow, which are becoming industry-standard.

Apple isn’t unique in producing a tool that attempts to take as much of the coding grind out of the machine-learning process as possible. For example, Google has AutoML, which is similarly designed to produce working machine-learning models with a minimum of code; there’s also Microsoft’s Machine Learning Studio, which attempts to boil down ML model building to a drag-and-drop process. Automating ML and A.I. is key to these technologies going as mainstream as possible.

The revelation of Overton is also interesting, as it shows that Apple’s researchers are moving on parallel tracks to other tech firms. Apple’s tool might help its internal staffers catch up to their rivals in A.I./ML, but it’s an open question whether they’ll ever transform it into a public-facing product, just as they’ve done for CoreML and other tools.

The post Apple ‘Overton’: Automating Low-Code Machine Learning appeared first on Artificial Intelligence.

Deep learning is everywhere. The catchphrase that powers today’s world of technology has immense implications — from teaching robots how to bluff, to learning complex musical construction and automating human-level language processing. In fact, artificial intelligence and machine learning are to our generation what the silicon chip was back in the ‘70s. However, while all that is great, there is a blind spot that Ryan Welsh spotted, one which had a rubber band effect on deep learning.

Explaining the advanced

It is this that gave birth to Kyndi, and an evolved deep learning model that Welsh and his team called ‘Explainable AI’. The technology, as Welsh explains in a conversation with News18, is about taking deep learning and AI algorithms from the ‘data’ stage to the ‘knowledge’ stage. How this differs is actually very simple — as Welsh tells us, “Statistical machine learning techniques are good at learning from data, but are not very good at reasoning. Knowledge-based AI approaches are good at reasoning, but cannot learn from data. Our Explainable AI software combines the two, so you have a system that is good at learning from data, and also good at reasoning. Thus, you get superior data efficiency, generalisation, and explainability compared to just using deep learning.”

This can have significant implications in fields such as legal affairs, market research, business analysis and development, education, insurance, and so on. How Kyndi’s approach differs is that it takes the abilities of a deep learning model and turns it into a higher level of information processing. The best example for this can be found in the impact that Explainable AI can have on a business analyst’s role. They call it IPA, or Intelligent Process Automation (IPA).

What it does, and who it is for

As Welsh explains, “For IPA, the user is the business analyst that has to read, analyze, and synthesize data. The beneficiary of the IPA process is the manager or VP level employee who uses the output of the business analyst to track performance/progress of a business process. That is, the manual process of being an analyst (e.g., reading, analyzing, and synthesizing data) is automated.”

In fact, Welsh further revealed that the advanced cognitive abilities of Explainable AI can be further used in the niche area of automating arts such as writing, or Natural Language Generation (NLG) in technical terms. He says, “It can be used for extractive and abstractive NLG. Although we do not use it for that yet. We currently use it for natural language understanding and reading comprehension which is the opposite of NLG. This is taking an idea and generating the appropriate text to convey that idea, whereas NLU/Reading Comprehension is about taking the text and trying to understand what idea is trying to be conveyed.”

In effect, think of this as the technology that replaces your business development analyst, who so far used deep learning tools to analyse company and market data, and present you with structured data sets complete with action points that can be directly implementable. Instead, Kyndi’s Explainable AI will do this, while also processing data that has been left behind. As Welsh states, “We build systems that read. Unstructured data is not in a format, so it has to be transformed into that format or it has to read by a human. We’ve built machines that can read, so we use them instead of humans. So, instead of leveraging only 20 percent of your structured data, you can now leverage 100 percent, with the other 80 percent being unstructured.”

Thinking responsibly

However, with such powers, come certain responsibilities. This brings us to two very important issues that need to be addressed for Explainable AI — dealing with bias, and setting legal precedence. From the way Welsh sees it, the very model of Explainable AI actually helps resolve bias that may be created as a result of deep learning algos. “If a deep learning system reviewed 100 resumés and recommended 10 for interviews, the user has cannot ask the system why it chose those 10. Whereas if a human chose 10, the user could as the human why they chose the 10 and the human would give an explanation. If the human said they chose the 10 because they are all white males, then the user can judge for themselves whether or not that is the criteria they want to filter for. Most people will conclude that to be sexist, and won’t use the results. But with a deep learning system and no explanation, you have no way to understand why the system generated the results,” he says. It is this that Explainable AI, which essentially explains to you the methodology used, aims to change.

As for setting precedent through technology, Welsh says, “The system is presenting outputs to users who then use that information to make decisions. Thus, it will not make a decision that sets precedent, rather the human always makes the decision, and the system is just an amplifier of their productivity. The reason you give explanations is so the user can evaluate the output of the system and decide for themselves it that aligns with their beliefs and should be considered in their final decision.” It is important to maintain this difference, for handing robots the ability to judge a legal trial will require a whole world of regulatory upheaval, and call for tertiary technologies to control such elements from the aspects of bias.

The next frontier

Kyndi’s Explainable AI model holds unlimited potential for how it can change the way advanced technology is implemented today. While today’s technology helps process the data, and the rest of the process is either human-driven or preset through human conditions, Explainable AI aims to evolve this to the next step, wherein the human-level effort would lie only in implementing the knowledge. This can, for instance, drastically reduce medical insurance processing times, upgrade the quality of home-based assisted schooling, trim inefficiencies in big corporate houses, and so on.

In the long run, Welsh hints at the possibility of Explainable AI also being implemented to create written documents through machine-generated language, or generally improve the quality of deep learning techniques used in areas such as fact checking through social media-spread misinformation campaigns. As Welsh states, “It is a knowledge revolution”.

The post Evolving Deep Learning: The Implications of Kyndi and its Explainable AI Technology appeared first on Artificial Intelligence.