Natural language processing with open source
Language is at the heart of everything we—humans—do. It’s now even more critical to attaining an advanced understanding of language in artificial intelligence (AI) when the goal is to achieve human-like intelligence in AI. Machines are learning to read, understand, and reason about data as humans, but we still have a lot of ground to cover, and this goal becomes a necessity as language pervades into multiple domains of AI from deep learning to robotics. Understanding language has been one of the most challenging tasks in AI due to its wide range of applications from voice assistants to communication among robots, and it remains to be so even after unprecedented improvements recently seen in AI (transfer learning, language models). The advancement in NLP and its subset natural language understanding (NLU) is not only an exciting prospect for technology but has become a necessity to overcome already existing and potential adversarial uses of this technology (such as deepfakes).
Fatma Tarlaci provides you with foundational knowledge in one of the most significant subfields of AI, NLP and NLU, through hands-on activities and teaches how to create an NLP model using open source NLP software, such as spaCy, natural language toolkit (NLTK), and PyTorch-NLP, to implement a variety of NLP tasks, such as question answering, sentiment analysis, and machine translation. By the end of the tutorial, you’ll be able to preprocess your data and implement your deep learning model to complete NLP tasks of your choice (sentiment analysis, question answering, machine translation, etc.).
Fatma reviews NLP, its tasks, and applications and details the current state of the field and the most widely used model architectures and concepts, such as neural networks, recurrent neural networks (RNNs), long short-term memories (LSTMs), transformers, transfer learning, and multitasking. You’ll explore text preprocessing, hyperparameter tuning, and model optimization. Once you finish reviewing, you’ll implement an NLP model architecture for an NLP task of your choice.
You’ll learn how to optimize and develop a better understanding of how the model implemented actually works, going over best practices for debugging deep learning models, and how to improve the performance of a model.
Relevant datasets and notebooks for the complete implementations of various sample models for each NLP task will be provided before you leave.
- A basic understanding of neural networks and deep learning
- Familiarity with Jupyter notebooks, Google Collaboratory, Python, and PyTorch
Materials or downloads needed in advance
- A laptop (Download the datasets Fatma provides.)
What you'll learn
- Understand NLP and NLU and the wide range of applications
- Experiment with end-to-end implementation of an NLP model
- Practice optimization and debugging techniques and PyTorch
- Learn about available open source software you can use for NLP
Fatma Tarlaci is a data science fellow and machine learning engineer at Quansight, where she focuses on creating trainings in AI and works closely on scientific computing within the open source community. She received her PhD in humanities before she transitioned into computer science, in which she received her master’s degree at Stanford University. Her work and research specializes in deep learning and data science.
Leave a Comment or Question
Help us make this conference the best it can be for you. Have questions you'd like this speaker to address? Suggestions for issues that deserve extra attention? Feedback that you'd like to share with the speaker and other attendees?
Join the conversation here (requires login)
Premier Diamond Sponsors
Premier Exhibitor Plus
For conference registration information and customer service
For more information on community discounts and trade opportunities with O’Reilly conferences
For information on exhibiting or sponsoring a conference
For media/analyst press inquires