> *"...when you have eliminated the impossible,
whatever remains, however improbable, must be the truth"*
-Sir Arthur Conan Doyle

Our brains process the meaning of a sentence like this rather quickly. 

We're able to surmise:
- Some things to be true: "You can find the right answer through the process of elimination.”
- Others that may have truth: "Ideas that are improbable are not impossible!"
- And some claims are clearly contradictory: "Things that you have ruled out as impossible are where the truth lies."

Natural language processing (NLP) has grown increasingly elaborate over the past few years. Machine learning models tackle question answering, text extraction, sentence generation, and many other complex tasks. But, can machines determine the relationships between sentences, or is that still left to humans? If NLP can be applied between sentences, this could have profound implications for fact-checking, identifying fake news, analyzing text, and much more. 


## The Challenge:

If you have two sentences, there are three ways they could be related: one could entail the other, one could contradict the other, or they could be unrelated. Natural Language Inferencing (NLI) is a popular NLP problem that involves determining how pairs of sentences (consisting of a premise and a hypothesis) are related. 

Your task is to create an NLI model that assigns labels of 0, 1, or 2 (corresponding to entailment, neutral, and contradiction) to pairs of premises and hypotheses. To make things more interesting, the train and test set include text in fifteen different languages! You can find more details on the dataset by reviewing the [Data page](https://www.kaggle.com/c/contradictory-my-dear-watson/data).

Today, the most common approaches to NLI problems include using embeddings and transformers like BERT. In this competition, we’re providing [a starter notebook](https://www.kaggle.com/anasofiauzsoy/tutorial-notebook) to try your hand at this problem using the power of Tensor Processing Units (TPUs). TPUs are powerful hardware accelerators specialized in deep learning tasks, including Natural Language Processing. Kaggle provides all users TPU Quota at no cost, which you can use to explore this competition. Check out our [TPU documentation](https://www.kaggle.com/docs/tpu) and [Kaggle’s YouTube playlist](https://www.youtube.com/playlist?list=PLqFaTIg4myu-1c3ygYzakW8-hNzQG59-5) for more information and resources.
Recommended Tutorial
We highly recommend
[this excellent tutorial on using KerasNLP to solve this problem, from the Keras team](https://www.kaggle.com/code/alexia/kerasnlp-starter-notebook-contradictory-dearwatson)
as well as
[Ana Sofia Uzsoy’s Tutorial](https://www.kaggle.com/anasofiauzsoy/tutorial-notebook)
that walks you through creating your very first submission step by step with TPUs and BERT.
This is a great opportunity to flex your NLP muscles and solve an exciting problem!

*Disclaimer: The dataset for this competition contains text that may be considered profane, vulgar, or offensive.*