Wipro – MEaningful Experiences Live 2017
No dia 9 de Novembro a Wipro, teve no seu evento MEaningful Experiences Live 2017 no Porto a presença do Vice-Presidente Global da Wipro, Mr. Murthy, ao qual aproveitamos para fazer umas perguntas que vos deixo abaixo transcritas.
Este evento focava no mercado do retalho as novas soluções inteligentes que estão em expansão, como a inteligência artificial e o machine learning. A parte mais interessante é ver que a tecnologia até já está bastante avançada, a nossa capacidade de a adaptar à realidade é que por vezes demora tempo com burocracias e leis prolongadas de formular.
# Who you think will “win” the autonomous vehicles segment? Do you think autonomous truck transport will hit the market before 2020?
From tech companies like Google, to car companies like GM, Ford, Toyota, to aggregators like Uber, 263 companies are directly or indirectly investing in or building autonomous vehicles. Volvo has tested autonomous trucks. While technology has matured on number of technical areas such as collision detection, speed control, threat perception, etc., it is the decision making in exception and safety scenarios that will take a long time to mature. This is because it requires a combination of human judgement, intuition and values which is difficult for a machine to emulate. This is in addition to a lot of tests, regulations and trial periods which the vehicles will have to go through. Hence, by 2020, we feel completely autonomous cars is a remote possibility.
# What will happen if one ML (machine learning) market analyzer fall into fraud data?
We understand that the question is about the Machine learning market analyzer data falling into wrong hands. In other words what will happen on data theft of the learning data of market analyzer. There are two types of data. First is the raw customer actions and transactions that are fed to machine learning analyzers. Second is the models and learning data that the ML systems generate and learn from. So, lets say if there is a breach:
1. If only learning data and models are compromised, then individual customers are still protected, because training data cannot be reverse engineered to identify individual customer inputs and actions. The learning data and models is of limited use to only generate actions for which the analyzer was designed for in first place.
2. However if customer actions and other input data is also compromised and has not been stored in encrypted or tokenized format, then it’s a serious breach. This can happen in regular transaction systems also. There is no ML aspect in this data breach.
# Will Artificial Intelligence change the way fashion and retail trends are set? Will AI develop in such way that robots will define trends by what they learn from people behaviour?
Partially. AI will definitely play a role in fashion, but in understanding fashion first and then predicting fashion trends. The beginning will be in analytics, such as capturing apparel images and identifying the upward trending designs, identifying patterns and materials that are getting popular. AI will also help suggest a particular fashion given a customer’s context such as body type, season, residential region. AI will then progress towards predictive analytics, where basis of the current trends and popularity, a machine learning algorithm can predict (with limited accuracy) on where the trend is progressing towards. If predictions done by AI systems turn out to be uniform and are taking seriously by fashion designers and makers; it is likely that AI will play a major role in influencing retail trends in near future.
Obrigado à Tábua Digital pelo convite para o evento