EqnMaster: Evaluating Mathematical Expressions with Generative Recurrent Networks
نویسندگان
چکیده
In this project, we seek to use computational models to infer mathematical structure from purely symbolic inputs. Though there are existing systems for parsing mathematical expressions, there have been only limited approaches to apply learning-based algorithms to this inherently subtle task. This project therefore proposes an integrated methodology that applies various configurations of neural networks to analyze sequences of mathematical language. We define both a discriminative task – in which we pursue the verification of a given equation – and a generative task – where we predict the evaluation of an input mathematical expression. We test our models on a novel dataset, and gain insight into the workings of our models by learning underlying representations of our data.
منابع مشابه
Learning a Generative Model for Validity in Complex Discrete Structures
Deep generative models have been successfully used to learn representations for high-dimensional discrete spaces by representing discrete objects as sequences, for which powerful sequence-based deep models can be employed. Unfortunately, these techniques are significantly hindered by the fact that these generative models often produce invalid sequences: sequences which do not represent any unde...
متن کاملBetter Generative Models for Sequential Data Problems: Bidirectional Recurrent Mixture Density Networks
This paper describes bidirectional recurrent mixture density networks, which can model multi-modal distributions of the type P(Xt Iyf) and P(Xt lXI, X2 , ... ,Xt-l, yf) without any explicit assumptions about the use of context . These expressions occur frequently in pattern recognition problems with sequential data, for example in speech recognition. Experiments show that the proposed generativ...
متن کاملA Variational Inequality Perspective on Generative Adversarial Nets
Stability has been a recurrent issue in training generative adversarial networks (GANs). One common way to tackle this issue has been to propose new formulations of the GAN objective. Yet, surprisingly few studies have looked at optimization methods specifically designed for this adversarial training. In this work, we review the “variational inequality” framework which contains most formulation...
متن کاملGenome Dreaming
We investigate and build the framework required to generate or “dream” of genomes that encode novel organisms. Importantly, we find that developing methods to validate the generated sequences is as important as building a good model to realize this task due to the lack of methods capable of interpreting and evaluating a generated sequence. We report on our results in building and optimizing a g...
متن کاملEvaluating Generative Models for Text Generation
Generating human quality text is a challenging problem because of ambiguity of meaning and difficulty in modeling long term semantic connections. Recurrent Neural Networks (RNNs) have shown promising results in this problem domain, with the most common approach to its training being to maximize the log predictive likelihood of each true token in the training sequence given the previously observ...
متن کامل