Wolfram Computation Meets Knowledge

Wolfram Summer School

Alumni

Jackreece Ejini

Science and Technology

Class of 2017

Bio

Jackreece Ejini is a self-taught computer programmer who got interested in the problem of making computers intelligent after watching the movie The Matrix in 1999. He studied genetics and biotechnology at the University of Calabar in Nigeria because he was seeking ways to shorten the growth and maturation of food crops to improve food security in his country. His other, secret reason for studying genetics and biotechnology was because he had been briefly exposed to genetic algorithms and genetic programming prior to gaining admission and he wanted to understand more of what was going on in the nucleus of a cell with the hope of applying it to the problem of making computers more intelligent.

After school, he continued his exploration into AI while also bettering his math and programming skills. While doing so, he came upon expert systems, GAs, GPs and a host of other AI paradigms, including artificial neural networks. He personally disliked artificial neural networks just because they were called neural networks, and he knew that brain neurons were more complex. He thought the “neural” part of the name was a marketing ploy and grossly misleading because he just couldn’t help thinking of the complexities of a real neuron any time he heard the name mentioned.

As he protested for years against neural networks, they got better and better—which forced him to investigate more about their functioning. Although they were good for pattern recognition and some people were already prophesying that they would one day overwhelm humanity in a bleak dystopian future, he still thought they were hacky, inelegant and kind of very slow, and they failed badly when they did—but he has now accepted them the way they are, as kind of a black box that works.

So he went on further in his investigations and met Stephen Wolfram’s A New Kind of Science—finally, hope on the horizon. He strongly believes that he will find some algorithm that simulates human thinking by searching the computational universe, either in systems already explored in the book, like cellular automata and network systems, etc. or in a newly invented system that follows simple rules. On the side, he is researching building apps that exploit most of the research done in sentic computing and also building a blockchain-backed electronic voter system.

Computational Essay

Humidity in the United States »

Project: Emotion Categorization of Text

Goal of the project:

To build an emotion categorization system that receives free-form text input from users and outputs a list of emotions expressed in the text.

Summary of work:

The work is based on the hourglass model of human emotions that classifies human emotions along the four dimensions of pleasantness, sensitivity, attention and aptitude, each of which have six levels (called sentic levels), amounting to a total of 24 emotions. We take data from the sentic net project and use it to build a pattern matcher that scores tweets based on the pre-modeled sentic net data. We then use the scored tweet data to train four classifiers along the individual dimensions using the Markov method.

Results and future work:

Using only 320,000 rows of analyzed and scored tweets out of 2,360,155 from the HC Corpus to train the classifiers (due to the large amount of time it takes for a highly efficient pattern matcher to process the individual tweets and the time constraints on the project), we were able to produce a usable web form that accepts input and produces some output that sometimes tends to be reasonable. Emotion categorization is an open-ended and difficult field, and our effort here is to see which methods and models deserve the most extensive effort in future work. In the future, we hope to train the classifiers on a much larger corpus of tweets, blog posts, news articles and other forms of user-generated content. If the results are still not satisfactory, we will explore other models of human emotions, such as the circumplex model, the vector model, the positive activation–negative activation (PANA) model, Plutchik’s model, the PAD (Pleasure, Arousal and Dominance) emotional state model and the Lövheim cube of emotion.