is an ever-growing, revolutionary field of science and technology. It is arguably the next great leap for society to reach a new and profound level of advancement. Artificial Intelligence will effect our daily social interactions, the way we cure diseases, pursue interstellar travel, analyze our brains, and figure out just what kind of cheeseburger will most satisfy us at this very moment. Nevertheless, artificial intelligence is only ever as good as its creators [well, this is an arguable statement according to the performance of machine learning projects in the last years].
Behind all the machine learning and data are the subtle and intricate designs of a team that has carefully put together a new kind of computer brain. Firms and companies over the world have the power and resources to make massive projects with server farms and the latest computing technology to advance their AI goals. However, this can only go so far. It's a matter of working smarter not harder. If the way your machine learns and processes information is fundamentally flawed or sloppily designed, all the resources in the world won't compensate for it. True professionals, who are able and willing to tweak and alter the algorithms and fundamentals to make it more efficient, are far more worth investment. The giant company that pumps their flawed designs with more computing power are doomed to only see their mistakes on a greater scale. Consider this most recent World Cup tournament.
AI researchers at Goldman Sachs, German Technische University of Dortmund
, Electronic Arts, Perm State National Research University, and others all ran machine learning models to predict the outcomes of several matches including the finals. The grand majority of them, despite all their resources, came up short with only Electronic Arts correctly predicting France's championship. With hindsight however, this makes sense. EA is video game company
famous for their soccer games. Their data, unlike their competition, includes the most up to date ratings for players. Furthermore, their techniques for machine learning were designed to add as much realism to player performance as possible. Their designs take into account the smallest but no less essential factors, all specifically designed to understand every variable of a soccer match. Their specialization and expertise was able to to easily surpass the incredible power and resources of firms like Goldman Sachs. Sometimes the mistakes are far more glaring. Google photos
has the ability to automatically detect images with the same background and fuse them together into one panoramic image. A Reddit User, 'MalletsDarker' attempted to use this machine learning feature by making Google images process three photos taken during a ski trip, two were landscapes and one was of his friend. Google images confidently merged the photos to create a breathtaking panorama of the ski resort. However, between two majestic mountains was the enormous profile of his friendis head peering out of the woods, bigger then anything around him. It's revealed that while google had created an impressive algorithm for merging images, it failed to take into account the most fundamental basics of picture composition.
Raw computing power and near infinite resources to grow are nothing to shake a stick at. However, if you don't have the experts who are willing to teak the intricacies of their AI, the results will always fall short.
Show All Articles