Reports from the Heart of the Machine
Editor: Du Wei
If you want to take the career path of deep learning, try this 401-page free face-to-face experience.
At present, deep learning has become the main driving force driving the development of artificial intelligence, convolutional neural networks, deep confidence networks, recurrent neural networks, etc. have been widely used in computer vision, speech recognition, natural language processing, audio recognition and bioinformatics and other fields, and obtain excellent results.
Learning deep learning has also become a hot topic, a variety of deep learning courses and books continue to emerge, the more famous in China is Li Mu Dashen's "Hands-on Learning Deep Learning".
This article is about the second edition of the book "Deep Learning Interviews", which is co-authored by two foreign authors, which contains hundreds of completely solved problems and covers a series of core topics in the field of AI.
This book is designed to provide rehearsal exercises for deep learning interviews or specific exams, as well as a logically rigorous overview of the field of deep learning for master's or PhD in machine learning and those waiting for an interview. The problems mentioned in the book are difficult to solve, but learned to significantly improve the reader's skills.
Book download address: https://arxiv.org/ftp/arxiv/papers/2201/2201.00650.pdf
This book is very valuable to students and job seekers who are able to confidently and fluently talk about any deep learning-related topic, answer technical questions clearly and correctly, and fully understand the purpose and meaning of interview questions. These will all be a huge advantage for them when it comes to interviews.
In terms of book content, this book covers a large list of topics related to deep learning job interviews and graduate-level exams. This also keeps the book at the forefront of scientific trends, teaching readers a core set of practical mathematical and computational skills.
What can I learn from this book?
Start your career
If you're a professional who is active in deep learning and data science, or want to do so, you're in luck. From deep learning to artificial intelligence, the labor market demand in various related fields is very high. Deep learning professionals are highly sought after and have become the highest paid workforce in companies around the world.
So, your career choice is the right one, and being rooted in a stable job can be of great benefit to your own financial and intellectual strength. However, it should be noted that the barriers to entry for these occupations are high and require deep learning interviews. In this type of interview, HR must also be adept at domain-specific topics to distinguish accurate and adequate interviewers from those who have little knowledge of the relevant applications.
Outside of the interview, the difference between the interviewers isn't always that important. Deep learning libraries are now so complete that they require very little high-skilled input to assemble their own machine learning pipelines. This level of competence doesn't do much to the interview. You'll be asked more practical, technical, and theoretical questions, and the interviewer expects you to answer all the questions confidently and fluently.
For unprepared interviewers, failing an interview means the end of a deep learning career path that many people will abandon after being rejected.
Advance your career
But if you've been working for years and are proactive, have superior computing power, and are ready to take on a more active and hands-on role in deep learning projects. You also have extensive knowledge in applied mathematics, computer science, statistics, and economics. All of these are great advantages that you have.
Even then, you can't say you're fully prepared for a deep learning interview, especially in the most interesting, autonomous, and challenging roles that require you not only to understand how to get the job done, but also to be able to demonstrate that knowledge clearly, confidently, and without hesitation. Some questions are straightforward and common during an interview, but others may be beyond the realm you've been exposed to since college.
There's no reason for that to happen, and when the interview comes, make sure you're ready. You need to know the latest terminology, concepts, and algorithms. You'll also need to update the basics in your memory and how you can apply them to your current research practices.
Delve into deep learning
The chapters of the book are structured as follows:
Introduction to topics;
Explain the problems that are at the heart of the topic;
Complete solutions.
The questions listed in each chapter are concise, practical, and relevant to the topic.
Problems are mainly divided into two categories, namely conceptual problems and applied problems. Conceptual questions are designed to practice or put into practice what you've learned (mostly related to Python and PyTorch).
The following figure shows an example of a conceptual problem in information theory:
The following figure shows an example of an application issue in PyTorch:
About the Author
Book author Shlomo Kashani, head of AI at DeepOncology, has developed deep learning techniques for precision tumor detection that expand and advance the capabilities of human experts. The CNN-dependent work marks the culmination of his career in applying AI technology to solving medical problems. Previously, he obtained a Master of Science (Honours) in Digital Signal Processing from the University of London.
Amir Ivry, editor-in-chief of the book, is a phD student at technion's School of Electrical and Computer Engineering and has published more than a dozen papers in IEEE core journals and top academic conferences. Since 2015, he has been an applied research scientist in the fields of deep learning and speech signal processing.
The full table of contents of the book is as follows:
Quickly build an enterprise-grade TTS speech synthesis assistant with NVIDIA Riva
NVIDIA Riva is an SDK that uses GPU acceleration to rapidly deploy high-performance conversational AI services for rapid development of speech AI applications. Riva is designed to help you easily and quickly access sessionAL AI capabilities, out of the box, and quickly build high-level TTS speech synthesis services with a few simple commands and API operations.
January 12, 2022, 19:30-21:00, the main introduction of this online sharing:
Introduction to speech synthesis
Introduction to NVIDIA Riva features
Launch the NVIDIA Riva client to quickly implement text-to-speech functionality
Use Python to quickly build Riva-based TTS speech synthesis service applications