Members:
- Ayush Desai - desai.ayus@northeastern.edu
- Harsh Shah - shah.harsh8@northeastern.edu
Problem description:
What problem are we solving?
- The problem we aim to address is the efficient translation of text from one language to another, leveraging computatinal methods. The objective is to maximize translation accuracy along with minimizing the loss and achieving balance between translation quality. The loss functions that will be used to train our model is sparse categorical cross entropy.
Formal description of the problem.
- Formally, text translation computationally will involve converting a sequence of symbols from a source language S (represented as X = x1, x2, ..., xn)) into an equivalent sequence in a target language T (represented as Y = (y1, y2, ..., ym)). The objective is to find a mapping ( f: X ⇒ Y) that accurately translates the input text into the target language maintaining the meaning and semantic structure of the output sentence.
What are the inputs and outputs?
- The inputs include the source text and the desired target language, while the output is the translated text in the specified language.
What data are we using?
- https://www.kaggle.com/datasets/devicharith/language-translation-englishfrench
- https://www.kaggle.com/datasets/kaushal2896/english-to-german
Why is it interesting?
- This problem is significant as it addresses the growing need for seamless communication across languages, facilitating global interaction, knowledge and cultural exchange.