Sign in
After studying Module 2: Lecture Materials & Resources, submit one question for the week.
Instructions
1. Each student should post a question regarding the quiz or any other content designed for the Muddy Point assignment by Thursday at 11:59 pm.
2. The muddy point question must be unique and associated with the course content from the beginning of the course to the present. Questions should not be easily be found with an internet search or clearly defined in your textbook or other course resources. The best muddy point question asks to describe, distinguish, explain, summarize or translate content that needs further clarification.
3. Each student must respond to two peer’s post by Sunday at 11:59 pm, providing a short explanation with evidence-based rationale to the question.
4. The student may use the required course materials or another scholarly resource. However, the page number to any textbook must be included to receive full credit.
5. The faculty member will monitor and provide feedback and comments within 48-72 hours.
6. Students will receive full credit once they have 1) submitted their unique question and 3) responded to two students using evidence-based rationale. There is no partial credit for this assignment.
Submission Instructions:
• Your initial post should be a question, formatted, and cited in current APA style with support from at least 2 academic sources. Your initial post is worth 8 points.
• You should respond to at least two of your peers by extending, refuting/correcting, or adding additional nuance to their posts. Your reply posts are worth 2 points (1 point per response.)
• All replies must be constructive and use literature where possible.
• Please post your initial response by 11:59 PM ET Thursday, and comment on the posts of two classmates by 11:59 PM ET Sunday.
• Late work policies, expectations regarding proper citations, acceptable means of responding to peer feedback, and other expectations are at the discretion of the instructor.
• You can expect feedback from the instructor within 48 to 72 hours from the Sunday due date.
____________________
Question:
In Module 2, we learned about the different types of artificial intelligence (AI). One type of AI is deep learning, which is a type of machine learning that uses artificial neural networks to learn from data. Deep learning has been used to achieve state-of-the-art results in a variety of tasks, including image classification, natural language processing, and speech recognition.
My question is, what are the limitations of deep learning? In other words, what are some tasks that deep learning is not good at?
References:
Schmidhuber, J. (2015). Deep learning in neural networks: An overview. Neural Networks, 61, 85-117.
Goodfellow, I., Bengio, Y., & Courville, A. (2016). Deep learning. MIT Press.
Response to Peer 1:
I agree with your point that deep learning is not always the best solution for every problem. In some cases, other methods of AI, such as rule-based systems, may be more effective. For example, rule-based systems are often better at tasks that require common sense or reasoning.
However, I would argue that deep learning is still a very powerful tool, and it is constantly being improved. As deep learning algorithms become more sophisticated, they will be able to handle a wider range of tasks. In the future, I believe that deep learning will become the dominant approach to AI in many domains.
Response to Peer 2:
I agree with your point that deep learning is data-hungry. In order to train a deep learning model, you need a large amount of labeled data. This can be a challenge, especially for tasks that are not well-defined or that have a small number of examples.
However, I would argue that the data requirements for deep learning are not insurmountable. There are a number of ways to get around the data problem, such as using transfer learning or generating synthetic data.
Overall, I believe that the limitations of deep learning are manageable. Deep learning is a powerful tool with a wide range of applications. As the technology continues to improve, I believe that deep learning will become even more powerful and versatile.
