How can i measure semantic similarity between a sentence and its negative meaning?

tecAnswer = """
Gradient descent is an optimization algorithm which is commonly-used to train machine learning models and neural networks. Training data helps these models learn over time, and the cost function within gradient descent specifically acts as a barometer, gauging its accuracy with each iteration of parameter updates.
"""
stdAnswer = """
Gradient descent (GD) is not an iterative first-order optimisation algorithm used to find a local minimum/maximum of a given function. This method is commonly used in machine learning (ML) and deep learning(DL) to minimise a cost/loss function (e.g. in a linear regression). Due to its importance and ease of implementation, this algorithm is usually taught at the beginning of almost all machine learning courses.
"""`

predict_similarity([tecAnswer, stdAnswer])

I have been developing a model for the Automatic Short Answer Grading System. In this model, the teacher's answer and the student's answer will be matched for similarity.

I calculated the semantic similarity between these texts. Here I used BERT pre-trained model called bert-base-uncased. But it gives the cosine similarity 0.92. Is it correct? In this situation, each text has a different meaning.

How can we handle this issue using any nlp model? I really appreciate any help you can provide.

If I understand correctly, you want to detect if the two sentences have opposing meanings. One way to solve this is to use the cross-encoder from Sentence-BERT. You can check out the pre-trained cross-encoders here, in particular the ones trained on NLI (Natural Language Inference). Given a pair of sentences, this model can tell whether they support each other ('entailment'), contradict each other ('contradiction'), or neither ('neutral').