In this paper, an evaluation of the performance of various CLIP models as zero shot face recognizers is done. The findings show that CLIP models perform well on face recognition tasks, but increasing the size of the CLIP model does not necessarily lead to improved accuracy. Additionally, there is an investigation of the robustness of CLIP models against data poisoning attacks by testing their performance on poisoned data. Through this analysis, the main aim was to understand the potential consequences and misuse of search engines built using CLIP models, which could potentially function as unintentional face recognition engines.

This project aims to predict the outcome of IPL matches using probability and machine learning. The project uses Big Data technologies such as MapReduce, Spark MLLib, and clustering in Scala. The data collection phase involved downloading and converting data from a website into a usable format, and then calculating runs and wickets for each batsman-bowler pair. Clustering was used to group similar players and make predictions for unknown pairs. The prediction phase uses probability to determine the outcome of each ball, and a machine learning approach using Spark MLLib and decision trees was used to improve the accuracy. The overall efficiency of the project was quite high, accurately predicting 8 out of 10 results.