En tiempos en que Sputnik todavía se presentaba como una alternativa a Selecciones del Reader’s Digest, mi viejo me regaló uno de mis primeros textos de tecnología: «El ABC de la Cibernética», de V. Kasatkin, lectura ideal para el ómnibus si ya habías completado una novela y querías cambiar un poco de aire. Quiero decir: números. Empezaba a probarme calculando, y festejaba mis aciertos de cuentas en hexadecimal verificando en la recién llegada calculadora científica, tosco ejercicio de aprendizaje por refuerzo.
Between stories of star travel and the conquests of mankind, other rather para-scientific texts predicted that in the nineties regular trips in helium zeppelins would take place, that they would show us the Amazon from above, and that by then man would be building his first apartments on Mars. It took me a few years, but I saw that it would not happen. Thus, I began to anticipate that some prophecies would come true, and others definitely never would(semi-supervised classification).
In a computer store, someone noticed that whoever bought floppy disks always carried a certain accessory that started to become popular, that is to say, this gentleman inferred a rule of association. Thus, his intuition led him to offer an innovative combo for the time: ten floppy disks plus the capacity doubler, something that today would suggest any recommendation system under these hypotheses.
Those hobbies turned into interests were outlining a possible direction for me as a student. Faced with the questions posed by such a new career, I observed(sentiment analysis) that my mother, worried, had some doubts about its feasibility. I could understand(topic extraction) that she was suggesting me to prepare myself for an alternative path, in order to compete in some call for a predictably stable job(time series analysis) and of notorious security(clustering). But I didn't change my mind.
They say that a couple of years ago they saw McFly's DeLorean go by. I don't know if I believe it, because before that, I had watched all the daily news in 1987 and I couldn't find a single news item about Buck Rogers going into space. Maybe my virtual assistant(speech to text) can shed some light on this; I'm going to ask him(question answering).
Non-rigorous glossary:
- Aprendizaje por refuerzo: incorporar en un algoritmo la maximización de alguna «recompensa» expresada como función objetivo
- Semi-supervised classification: classification algorithm whose training uses some unlabeled items, and many labeled ones.
- Association rule: pattern of association between certain attribute values - for example, detecting that certain values occur at the same time.
- Recommender system: algorithm that suggests a next option or decision by trying to predict the rating of that option or decision.
- Sentiment analysis: polarity (positive or negative) of an expression in natural language, with respect to a frame of reference.
- Topic extraction: in a natural language text, identify the topic(s) addressed, with a degree of certainty.
- Time series analysis: data mining techniques oriented to locate patterns and elaborate predictive models on data that present periodicity characteristics.
- Clustering: identification of the best groups according to which a data set can be classified.
- Speech to text: from the audio signal of the voice, transcribe the text it represents.
- Question answering: ideal system capable of understanding and answering, in natural language, questions on a given subject area.
Fernando López Bello @fer_lopezbello
Computer Engineer, PMP
Big Data Specialist
*This article is part of the Quanam Blog: more information
Connect