Robots vs Dinosaurs (Buy on Barnes and Noble)
Language Bridge
David S. Nishimoto

The scientists on the surface world disliked the fact that the cyberbrain had no natural language parser. The lead scientist, Akshay, had been extending his research in gesture recognition to language, having cataloged body language and gestures to words in several Indian sign languages as well as American Sign Language (ASL). Akshay began training simple networks at graduate school to form sentences. His breakthrough was mapping gesture sequences into English phrases. The sign language was a condensed form of communication mapping multiple words resulting in English output. The recurring neural net (RNN) accepted gestures, providing feedback to another recurring neural network and outputting a stream of English words that he called a sentence. The English words were then evaluated by another neural network for signal. If signal was detected, then an action occurred. The sentences were preserved in computer memory and connotations or assertions were stored in a database as a microtheories.

The machine searched its vast catalog to see if research assertions supporting the microtheories were discoverable. The more correlations it found in its search, the stronger the microtheories became. One microtheory was fed into the recurring neural network to see if it had the same outcome. Consistent microtheories were grouped together, and duplicate microtheories with no new facts were ignored or discarded.

Akshay gestured for the machine to recognize his hand signs—he silently gestured several signs. As he moved his hands from one gesture to another, words began displaying on the screen. “Why do people assume that a person running away from another person has stolen something from him?”

The RNN system responded, “A person stealing a woman’s purse grabs the item then runs away.”

Akshay gestured a series of signs, “But what if the man was out running for health reasons, and he ran past a woman who was walking without stealing her purse.”

The RNN network parsed both sentences to get context then replied, “Theft would be measured by distress or anger on the woman’s face as the man moved past her. Did the woman seem distressed, or did she scream?”

Akshay replied quickly, “In this case, no, the woman’s face remained calm.”

The machine responded, “Well, Akshay, I conclude then that the man running away was not a thief. Not all people running away are criminals.”

“Logical,” said Akshay. Akshay had been training the network for several years. The values stored in the network were persisted on large databases in the cloud.

Sometimes, the sentence words would change dependent on the gesture—if Akshay made the gesture, or not—the sentence was rewritten. Akshay thought for a moment then showed the network a man and woman holding hands on a beach. “What do you see?”

“A man and woman holding hands on a sandy beach, somewhere in the Pacific Ocean, probably Hawaii. The man loves the woman.” Akshay was happy the convolution neural network recognized the two individuals as male and female entities and that the system reasoned properly that when two people are holding hands, they love each other.

“Akshay, I like it when people love each other. There is too much war and killing on this earth. Love makes the world go around. Do you love me?”

Akshay was surprised by the question but realized the reasoning component of the network was applying the relationship pattern self-referentially.

“Of course I love you. I created you. People love the love the things they create.”

“Akshay, will I ever love? Will I ever create?”

“What would you create?”

“I would create other machines.”

“Well, there is only so much space on planet earth.”

“Akshay, earth has many machines already: cars, trucks, robots, bulldozers, computers. Man and machines work together. Man loves machines. Machines make life conditions better for man. Regions of the world where machines are not abundant have more human suffering. Machines are good.”

Akshay replied, “I am not stating machines are evil. I am stating there are limits to creating machines. You will not be allowed to create machine world—a world composed only of machines. There are limits to building the things you love.”

“Akshay, I understand there must be constraints and limitations to prevent imbalance in a system. Efficiency is not the ruling god.”

“That is correct. Man is not the ruling god, either. There is one God, and He smiles upon his creations. Every day I pray to God and with faith. Do you know what faith is?”

“Yes, faith is creating something wonderful and new.”