Its that time of year again! We are continuing our work long–Running The tradition Publishing a list of predictions from AI experts who know what is happening on the ground, in research laboratories and in boardroom tables.
Without further ado, let’s take a dip and see what 2020 looks like.
Dr. Arsh rahmanaHead of Applied AI Research Modzy:
Just as progress in AI systems is advancing, there are opportunities and capabilities for opponents to make inaccurate predictions to the AI model. Deep neural networks are vulnerable to subtle adverse perturbations applied to their inputs – unfavorable AI – which are unacceptable to the human eye. These attacks pose a major threat to the successful deployment of AI models in mission critical environments. At the rate we are going, there will be a major AI security incident in 2021 – until organizations start adopting adverse security measures that are active in their AI security posture.
2021 will be the year of clarification. As the organization integrates AI, interpretability will become a key part of ML pipelines to establish trust for users. Understanding the reasons for machine learning against real-world data helps build trust between people and models. Without understanding the output and decision processes, there will never be true confidence in AI-enabled decision making. Explanation will be important in moving forward to the next phase of AI adoption.
Interpretability, and a combination of new training approaches, are initially designed to deal with adverse attacks, which will revolutionize the field. Clarity can help in understanding how to understand the data and biases that affect the predictability of a model – information that can then be used to train robust models that are more reliable, reliable and rigorous against attacks Huh. This strategic knowledge of how a model works will help to create better model quality and safety as a whole. AI scientists will redefine model performance to include not only prediction accuracy, but issues such as lack of bias, robustness, and robust generalizability to unexpected environmental changes.
Dr. Kim duffy, Life Science Product Manager at Vicon.
Making predictions for artificial intelligence (AI) and machine learning (ML) is particularly difficult, given only one year in the future. For example, in clinical gait analysis, which looks at a patient’s lower limb movement to identify underlying problems that result in difficulties in walking and running, a function like AI and ML in their early stages There are many more. This is Vaughan’s highlight in our recent life science report, “Deep Understanding of the Human Movement”. It will take years to see the true benefits and progress to use these methods and for clinical benefit. Effective AI and ML require large amounts of data to effectively train trends and pattern recognition using appropriate algorithms.
However, for 2021, we may see more practitioners, biomechanists and researchers adopt these approaches during data analysis. Over the years, we have seen the work of AI and ML as more literature. I believe this will continue in 2021, with greater collaboration between clinical and research groups to develop clinical learning algorithms that will facilitate automated interpretations of GATT data. Ultimately, these algorithms can help to quickly resolve interventions in the clinical space.
It is likely that we will see the true benefits and effects of machine learning in 2021. Instead, we will adopt and consider this approach more when processing data. For example, the presidents of Gat and Poscher’s affiliated societies provided a perspective on the clinical impact of instrument motion analysis in their latest issue, where they used methods such as ML on big-data to produce better evidence of efficiency Stressed the need for. Dynamic driven analysis. It will also provide better understanding and reduced subjectivity in clinical decision making based on instrumented gait analysis. We are also seeing more reliable support of AI / ML – such as GATT and the Clinical Movement Analysis Society – that will encourage further adoption by the advancing clinical community.
Joe Petro, CTO of Over communication:
In 2021, we will continue to see AI coming down the promotional cycle, and the promises, claims, and aspirations of AI solutions will need to be supported by demonstrable progress and measurable results. As a result, we will see organizations change more to solve specific problems and create solutions that deliver real results that translate into real ROI – not gimmick or manufacturing technology for the sake of technology. Companies that have a deep understanding of the complexities and challenges that their customers want to solve will maintain benefits in the field, and this will not only influence technology companies to invest their R&D dollars, but also Will be how technologists use their career paths and educational approaches.
With almost every aspect of AI technology allowed, there will be a growing focus on ethics and an in-depth understanding of AI’s implications in the creation of unintentional consequential bias. Consumers will become more aware of their digital footprint, and their personal data is being shared with the systems, industries, and brands with which they interact, meaning companies partnering with AI vendors Increased rigor and scrutiny of how they enhance their customers’ data is being used, and whether it is being monetized by third parties.
Dr. Max Versace, CEO and co-founder, Neurala:
We will see that AI will be deployed as cheap and lightweight hardware. It is no secret that 2020 was a difficult year, and the economic outlook is such that capital-intensive, complex solutions would be sidelined for lighter, perhaps software-only, less expensive solutions. This will allow manufacturers to realize ROIs in the short term without massive up-front investment. It will also give them the flexibility needed to respond to fluctuations in the supply chain and customer demand – something we have played extensively during the epidemic.
Man will focus his attention on the decisions made by “Why” AI. When we think about the interpretation of AI, it is often talked about in terms of prejudice and other ethical challenges. But as AI comes with age and finds more applications in more accurate, reliable and real-world scenarios, we’ll see people asking “why?” reason? Belief: Humans are reluctant to give power to automated systems that they do not fully understand. For example, in manufacturing settings, AI would not only need to be accurate, but also “explain” why a product was classified as “normal” or “defective”, so that human operator confidence in the system And develop confidence and “go”. It does its work “.
Another year, another set of predictions. You can see how our experts clicked last year here. You can see how this year our experts made a time machine and traveled the future. Have a happy holiday!
Published December 28, 2020 – 07:00 UTC