She explained with practical examples the impact that A.I and deep learning is already having and more importantly the science behind it as;
- Artificial Intelligence – Enhancing knowledge, reason, planning and perceiving mechanisms.
- Machine Learning – Enhancing learning from data, expert systems and hand-crafted classifications.
- Deep Learning – Enhancing learning from data, neural networks and computer learned features.
Whether it be electrification, connectivity, shared mobility or autonomous driving, a combination of the above methods with the correct use of big data and Graphics Processing Units (GPU) acceleration will allow new technologies to be created that solve old hardware limitation problems, making GPU’s ready to assist in the transition towards ‘A.I. of IoT’. New data models are beginning to form, where the data required to enhance images can be done more confidently using handcrafted classification before being applied to particular machine learning models.
Any specimen object, like an image or sound, has its own form and frequency, which helps identify suitable subject analysis techniques such as e.g., spectrogram analysis, that help the object features to be placed into automatic weightings (supervised or unsupervised) and these techniques then remove the need for handcrafting classification as parameters are instead predicted. These efficiencies can then be applied into advanced techniques like neutral networks that need stronger input confidence giving rise to more use cases, such as biomedical imaging, and so, auto-encoding is creating many broader advances, within health, retail and education as it can be better programmed.
Taking us one step closer to realising A.I. for everyday use cases, this webinar explained how A.I. fundamentals can work together practically.