Min-Chih Liu

I have a dream, which is to participate in the development of robots with general artificial intelligence, similar to those in the movie 'Terminator'.

Therefore, I chose to study control algorithms in graduate school and selected a topic related to industrial robots for my thesis.

I subsequently pursued work in the related field, hoping to translate the theories I learned into practical results.

In recent years, with the leap in artificial intelligence development and the emergence of generative AI, people have started to foresee that AI will become part of everyday life, making the realization of general artificial intelligence no longer an unattainable dream.

To make up for my lack of knowledge in this area, I have invested resources in learning AI project development and implementation, from image recognition and speech model training to natural language/emotion monitoring and large language model API applications. I hope that one day I can get closer to my dream.

TQC+ Machine Learning Professional

1. Use Tensorflow Keras packages and VGG19 model to train Customized models and implement some mask detection, flower classification...etc.

2. Use PyTorch, Ultralytics packages and YOLOv8 model to detect object and integrated with Pysied6,Qtdesigner to implement a Windows application.

3. Use librosa package to transfer .wav file to mfcc and train some voice detection models and use nltk, gensim packages to implement word to vector process, finally train a setiment model using keras embedding and lstm layer.

4. Use nginx and certbot to build a static server and Use flask,line-bot-sdk and gemini api packages to implement a chat bot application.

5. Use lanchain, sentence-transformers, llama-cpp-python packaes to implement RAG application. By pre-feeding documents, large language models can generate targeted responses based on the content described in the documents to ensure that the generated content avoids being too vague.