Neural Generative Question Answering (GENQA) System

A Neural Generative Question Answering (GENQA) System is a QA System that is based on a Convolutional Neural Network, Recurrent Neural Network and a Multi-layer Perceptron training systems.

• Context:
  • It implements GENQA Alogrithms to solve a GENQA Task.
  • It composed of the following subsystems:
    • GENQA Interpreter System,
    • GENQA Enquirer System,
    • GENQA Answerer System.
• Example(s):
  • Question: “Which country does Xavi play for as a midfielder?” (translated from Chinese), GENQA system's Answer: “He plays for Spain”
• Counter-Example(s):
  • a Deep Similarity Neural Network QA System,
  • an Embedding-based QA System,
  • a Neural Responding Machine (NRM),
  • a Retrieval-based QA System.
• See: Artificial Neural Network, Deep Learning Neural Network, Natural Language Processing Task, Attention Mechanism, Long Short-Term Memory (LSTM) RNN Model.

---

---

References

2016

• (Yin et al., 2016) ⇒ Jun Yin, Xin Jiang, Zhengdong Lu, Lifeng Shang, Hang Li, and Xiaoming Li. (2016). “Neural Generative Question Answering.” In: Proceedings of the Twenty-Fifth International Joint Conference on Artificial Intelligence (IJCAI-16). ISBN:978-1-57735-770-4. arXiv preprint arXiv:1512.01337.
  • QUOTE: ...we propose an end-to-end neural network model for generative QA, named GENQA, which is illustrated in Figure 1.
    

    Figure 1: System diagram of GENQA.

    The GENQA model consists of Interpreter, Enquirer, Answerer, and an external knowledge-base. Answerer further consists of Attention Model and Generator. Basically, Interpreter transforms the natural language question Q into a representation H_Q and saves it in the short-term memory. Enquirer takes H_Q as input to interact with the knowledge-base in the long-term memory, retrieves relevant facts (triples) from the knowledge-base, and summarizes the result in a vector r_Q. The Answerer feeds on the question representation H_Q (through the Attention Model) as well as the vector r_Q and generates an answer with Generator. We elaborate each component hereafter.

---