Welcome to the staging ground for new communities! Each proposal has a description in the "Descriptions" category and a body of questions and answers in "Incubator Q&A". You can ask questions (and get answers, we hope!) right away, and start new proposals.
Are you here to participate in a specific proposal? Click on the proposal tag (with the dark outline) to see only posts about that proposal and not all of the others that are in progress. Tags are at the bottom of each post.
Activity for mr Tsjolderâ€
| Type | On... | Excerpt | Status | Date |
|---|---|---|---|---|
| Edit | Post #291536 | Initial revision | — | 6 months ago |
| Answer | — |
A: Do Large Language Models "reason"? I am afraid that the answer to your question is actually quite simple: > there is no consensus among scientists. This is illustrated by the fact that the "fathers of Deep Learning" do not entirely agree on the current state of AI: - Geoffrey Hinton strongly believes that human-level/human-lik... (more) |
— | 6 months ago |
| Comment | Post #289179 |
If you feel like a statistics community might make sense, you should definitely feel free to launch another proposal. After all, this proposal has not gained much traction yet and it might be that your proposal resonates with more people.
This being said, chances are high that there will be overla... (more) |
— | 6 months ago |
| Edit | Post #290190 | Initial revision | — | about 1 year ago |
| Answer | — |
A: What is the technical term for converting a sound recording to a phoneme vector? I am no specialist on audio processing, but I believe the general task of converting raw audio to discrete symbols is called transcription. However, this does not distinguish whether these symbols are phones, phonemes or text. I strongly suspect that there is no more specific terminology (as in N... (more) |
— | about 1 year ago |
| Edit | Post #290026 | Initial revision | — | about 1 year ago |
| Answer | — |
A: Is LLM fine tuning the same as transfer learning? TL;DR: Fine-tuning is not the same as transfer learning, but is often a part of transfer learning. The goal of transfer learning is to "transfer" information from one model to some other model. Typically, the goal is to reuse parts of a good-performing, existing model to create a new model (possi... (more) |
— | about 1 year ago |
| Comment | Post #290004 |
Does incremental training refer to [incremental learning](https://en.wikipedia.org/wiki/Incremental_learning)? Also, could you shortly explain why you think they would be the same? (more) |
— | about 1 year ago |
| Edit | Post #289179 |
Post edited: add section on minimum effort requirement |
— | about 1 year ago |
| Comment | Post #289983 |
I would argue that the linked question focuses on what paradigm to use in what context. Furthermore, the question (vaguely) acknowledges the existence of other "learning" terms, but you are definitely right that the list of approaches is a nice addition and I updated the question accordingly. (more) |
— | about 1 year ago |
| Edit | Post #289949 |
Post edited: add link to wikipedia list of approaches and add tag |
— | about 1 year ago |
| Edit | Post #290014 | Initial revision | — | about 1 year ago |
| Answer | — |
A: What is unsupervised learning? Unsupervised learning covers a variety of different tasks. Depending on the task at hand, different techniques can be used. However, there are a few common paradigms that are used to extract information from unlabelled data. A simple, yet effective method is to create labels from the input data.... (more) |
— | about 1 year ago |
| Comment | Post #289983 |
I would argue that the second sentence on the linked page gives a clear answer to the question. If you think it is too technical, then it would be good to include this information in your question. The problem is that my answer would end up being very similar to the information in this second sentenc... (more) |
— | about 1 year ago |
| Comment | Post #289981 |
I think this is a nice write-up. For me, it misses the point a bit on unsupervised learning and I would not use transfer learning to talk about incremental training (which I know best as continual learning). As a matter of fact, I would argue that fine-tuning, which is not specific to LLMs) is more c... (more) |
— | about 1 year ago |
| Comment | Post #289983 |
The page on hyper-parameters on [wikipedia](https://en.wikipedia.org/wiki/Hyperparameter_(machine_learning)) seems to answer this question pretty well. Do you find something confusing about the explanation there? (more) |
— | about 1 year ago |
| Edit | Post #289949 | Initial revision | — | about 1 year ago |
| Question | — |
What types of learning should I know about? In machine learning there seem to be various types of learning. I commonly hear about - supervised learning, - unsupervised learning, and - reinforcement learning. However, it is not always clear to me how they differ or what paradigm is best used for a particular task. I can also imagin... (more) |
— | about 1 year ago |
| Edit | Post #289713 | Initial revision | — | about 1 year ago |
| Answer | — |
A: Why can't we use linear regression on logits for solving logistic regression problems? There is the obvious answer that points out that since $y \in \{0, 1\}$, we would end up with invalid values for computing the logits. However, by transforming the labels to have values $\tilde{y} \in \{\varepsilon, 1 - \varepsilon\}$ for some $\varepsilon \in (0, 0.5)$ (also known as label smoothin... (more) |
— | about 1 year ago |
| Edit | Post #289179 |
Post edited: include prompt engineering, exclude "classical" AI |
— | over 1 year ago |
| Comment | Post #289268 |
I have had pretty good experiences with the markdown parsing from [mattermost](https://github.com/mattermost/mattermost). There might be edge cases, but they seem to be able to distinguish between dollar signs used for mathjax and as currency symbol. (more) |
— | over 1 year ago |
| Comment | Post #289196 |
There is definitely room for reconsidering the prompt engineering aspect for ML. However, the idea for the ML community would be to gather people who are interested in the technical aspects of ML, rather than people who just want to participate in the hype without caring for the technical side. Ther... (more) |
— | over 1 year ago |
| Edit | Post #289266 | Initial revision | — | over 1 year ago |
| Question | — |
Why can't we use linear regression on logits for solving logistic regression problems? Consider a machine learning problem with inputs $\boldsymbol{X} \in \mathbb{R}^{N \times D}$ and corresponding labels $\boldsymbol{y} \in \mathcal{Y}^N$. When the problem is to solve a regression task, $\mathcal{Y}^N = \mathbb{R}^N$ and we can use linear regression model, $$\hat{\boldsymbol{y}} =... (more) |
— | over 1 year ago |
| Edit | Post #289198 | Initial revision | — | over 1 year ago |
| Question | — |
Would it be possible to enable special features in the Incubator Q&A? Is it possible to have special features enabled in the Incubator Q&A? I have tried to port one of my an answers for a question on Machine Learning, but it was written for markdown with MathJax support. Without the MathJax support, the post is practically unreadable. Would it be possible to enabl... (more) |
— | over 1 year ago |
| Edit | Post #289197 | Initial revision | — | over 1 year ago |
| Answer | — |
A: What are the support vectors in a soft-margin SVM The goal of a soft-margin or C-SVM is to solve the following minimisation problem: $$\min{\boldsymbol{w}, b} \frac{1}{2} \|\boldsymbol{w}\|^2 + C \sumi \xii$$ subject to $$\forall i : \begin{aligned}yi (\boldsymbol{w} \cdot \boldsymbol{x}i + b) - 1 + \xii &\geq 0 \\ \xii &\geq 0\end{aligned},$$ ... (more) |
— | over 1 year ago |
| Comment | Post #289179 |
therefore it might make sense to set up a second section. Maybe it could be an explicit separation between technical and non-technical questions. However, I am not sure if it is a good idea to have both areas in the same community. Alternatively, we could make prompt-engineering on-topic, but it woul... (more) |
— | over 1 year ago |
| Edit | Post #289185 | Initial revision | — | over 1 year ago |
| Question | — |
What are the support vectors in a soft-margin SVM I know what Support Vector Machines (SVMs) are and how they work, but I regularly get confused by what exactly the support vectors are. In case of linearly separable data, the support vectors are those data points that lie (exactly) on the borders of the margins. After all, these are the only p... (more) |
— | over 1 year ago |
| Edit | Post #289179 | Initial revision | — | over 1 year ago |
| Answer | — |
Machine Learning The Machine Learning (ML) community aims to provide a platform for anyone interested in helping computers learn from data. We mainly target people with backgrounds in data science, machine learning and data visualisation, but also statisticians, computer scientists and anyone looking to learn about ... (more) |
— | over 1 year ago |
| Comment | Post #289124 |
As an ML researcher this feels like the proposal for a Q&A about ML tools (similar to how programming languages are the tools of programmers):
- latest models ~ latest programming language
- suitability of tasks ~ what language for what task?
- API usage ~ language syntax
- model deployme... (more) |
— | over 1 year ago |
| Edit | Post #289163 | Initial revision | — | over 1 year ago |
| Answer | — |
A: Free alternatives to GPT-3/4? Technically, chat-GPT (GPT-3.5) is also freely accessible via https://chat.openai.com Possible alternatives are: - TII's Falcon - LAION's Open Assistant - Mosaic's MPT - Google's Bard - Anthropic's Claude-2 - Meta's LLama 2 - ... This list is probably incomplete, especially if we ... (more) |
— | over 1 year ago |