Analytics Insight explains the unexpected challenges from Explainable AI in 2021.

The emergence of cutting-edge technologies has introduced another form of AI known as Explainable AI or XAI in the global market. It is a set of frameworks that help human users understand and trust the interpreted predictions and solutions from machine learning algorithms. The advancements of AI technologies are creating challenges for humans to comprehend the entire process of receiving specific outcomes from these machine learning algorithms. The black box models are created from real-time data that are making it impossible for humans to understand the calculation process. Sometimes the functionalities of ML models or neural networks are difficult to comprehend due to the complicated process. But it is essential for companies and start-ups to have a complete understanding of the rapid decision-making process. It is not often suggestible to blindly trust the AI models because their performance can change if there is a shift in the type of data or biased results based on the demographic and geographic segments. Thus, Explainable AI is the key requirement to promote end-users trust in large-scale implementation of AI models with appropriate explainability and accountability.

• EXPLAINABLE AI: MAKING DECISION-MAKING TRANSPARENT AND INNOVATIVE
• HOW DO WE CREATE TRUSTWORTHY AI WITH AI ETHICS AND TRANSPARENCY?
• MICROSOFT IS PUTTING AI TO WORK FOR A SUSTAINABLE PLANET
• SPOTLIGHT ON AI: LATEST DEVELOPMENTS IN THE FIELD OF ARTIFICIAL INTELLIGENCE

Explainable AI helps organizations to make the stakeholders understand the types of behaviors of AI models through monitoring model insights. There are multiple benefits of Explainable AI such as simplifying the complicated process of model evaluation, continuous monitoring and managing AI models to optimize business insights, and mitigating risks of unintended bias by keeping the models explainable and transparent. That being said, certain concerns with Explainable AI are rising too.

The first concern is the primary function of Explainable AI – explanation with transparency. This policy is becoming a threat for organisations that are continuously innovating new AI models or technologies with machine learning algorithms. The reason is that the creators have to explain and be transparent about the whole process and performance of the whole model to the stakeholders for a better understanding. The firms do not want to disclose all types of confidential information, trade secrets, and source codes to the public for security concerns. Then what will happen to the intellectual property rights that distinguish each company from one another? This is one of the unexpected challenges from the Explainable AI to innovators and entrepreneurs.

The second concern is that machine learning algorithms are highly complex and intangible in nature. Software developers or machine learning engineers can make common people understand the process of creating algorithms but the inner tangible process is very difficult to explain. Customers use these AI products subconsciously in their daily life such as face recognition locks, voice assistants, virtual reality headsets, and so on. But do they really need to know the complicated process in this fast-paced life? This information tends to become a little uninteresting and time-consuming to some stakeholders.

The third concern is for organizations to tackle different forms of explanation for different users with different contexts. Even if any company wants to follow the Explainable AI policy of making people understand the algorithms, different stakeholders can ask about different explanations such as technical details, functionalities, data management, factors affecting the result, and so on. The explanation should reflect the needs and wants of the stakeholders effectively for better stakeholder engagement. But sometimes it is impossible for organizations to answer so many questions at one time.

The fourth concern is receiving unreliable outcomes from these black boxes. Users should trust business insights from AI models, but it consists of potential risks. The system can generate misleading explanations due to a change in data. Then, the users will trust the error with utmost confidence that can lead to a massive failure in the market. These explanations are useful for the short term but not for the long-term plans.

That being said, despite the unexpected challenges from Explainable AI, companies can consider these five essential points to drive appropriate insights from AI models— monitor fairness and debiasing, analyze the models to drift mitigation, apply model risk management, explain the dependencies of machine learning algorithms as well as deploy the projects across different types of clouds.

The post EXPECT THE UNEXPECTED FROM EXPLAINABLE AI IN THE 21ST CENTURY appeared first on Artificial Intelligence.

Researchers from the University of Toronto and LG AI Research have developed an “explainable” artificial intelligence (XAI) algorithm that can help identify and eliminate defects in display screens.

The new algorithm, which outperformed comparable approaches on industry benchmarks, was developed through an ongoing AI research collaboration between LG and U of T that was expanded in 2019 with a focus on AI applications for businesses.

Researchers say the XAI algorithm could potentially be applied in other fields that require a window into how machine learning makes its decisions, including the interpretation of data from medical scans.

XAI is an emerging field that addresses issues with the ‘black box’ approach of machine learning strategies.

In a black-box model, a computer might be given a set of training data in the form of millions of labeled images. By analyzing the data, the algorithm learns to associate certain features of the input (images) with certain outputs (labels). Eventually, it can correctly attach labels to images it has never seen before.

The machine decides for itself which aspects of the image to pay attention to and which to ignore, meaning its designers will never know exactly how it arrives at a result.

But such a “black box” model presents challenges when it’s applied to areas such as health care, law, and insurance.

For example, a [machine learning] model might determine a patient has a 90 percent chance of having a tumor. The consequences of acting on inaccurate or biased information are literally life or death. To fully understand and interpret the model’s prediction, the doctor needs to know how the algorithm arrived at it.In contrast to traditional machine learning, XAI is designed to be a “glass box” approach that makes decision-making transparent. XAI algorithms are run simultaneously with traditional algorithms to audit the validity and the level of their learning performance. The approach also provides opportunities to carry out debugging and find training efficiencies.

The first, known as backpropagation, relies on the underlying AI architecture to quickly calculate how the network’s prediction corresponds to its input. The second, known as a perturbation, sacrifice some speed for accuracy and involves changing data inputs and tracking the corresponding outputs to determine the necessary compensation.

There is a lot of potential in SISE for widespread application. The problem and intent of the particular scenario will always require adjustments to the algorithm—but these heat maps or ‘explanation maps’ could be more easily interpreted by, for example, a medical professional.

LG’s goal in partnering with the University of Toronto is to become a world leader in AI innovation. This first achievement in XAI speaks to our company’s ongoing efforts to make contributions in multiple areas, such as the functionality of LG products, innovation of manufacturing, management of supply chain, the efficiency of material discovery, and others, using AI to enhance customer satisfaction.

When both sets of researchers come to the table with their respective points of view, it can often accelerate problem-solving. It is invaluable for graduate students to be exposed to this process.

While it was a challenge for the team to meet the aggressive accuracy and run-time targets within the year-long project—all while juggling Toronto/Seoul time zones and working under COVID-19 constraints—Sudhakar says the opportunity to generate a practical solution for a world-renowned manufacturer was well worth the effort.

The post This New Algorithm can Explain Artificial Intelligence (XAI) appeared first on Artificial Intelligence.