13

What you are looking for is called "reinforcement learning". A reinforcement learning algorithm will try to maximize a reward function. This reward represents how "good" or "bad" an action is in the actual context. For example, in the snake game, your reward will be positive for eating an apple and negative when the snake hits a ...


12

Active learning (AL) is a weakly supervised learning (WSL) technique where you can have both labelled and unlabelled data [1]. The main idea behind AL is that the learner (or learning algorithm) can query an "oracle" (e.g. a human) to label some unlabelled instances. AL is similar to semi-supervised learning (SSL), which is also a WSL technique, ...


9

MNIST (along with CIFAR) may be the "Hello World" of supervised learning for image classification, but it is definitely not the "Hello World" of all machine learning techniques, given that RL is also part of ML and MNIST is definitely not the "Hello World" of RL. I don't think there is a single "Hello World" problem ...


7

In the context of ML, ground truth refers to information provided by direct observation (empirical evidence). If you're training an algorithm to classify your data, then the ground truth will be the actual, true labels which could for example be manually annotated by an domain expert. Please note that the models prediction or the inferred labels, are not ...


6

As @desertnaut mentioned in the comment No weak learner becomes strong; it is the ensemble of the weak learners that turns out to be strong Boosting is an ensemble method that integrates multiple models(called as weak learners) to produce a supermodel (Strong learner). Basically boosting is to train weak learners sequentially, each trying to correct its ...


6

The probability density is used to 'measure how good' the parameters are because it is a natural way of quantifying if these parameters are good for the observed data. Also, as the notation often causes some confusion, $L(\theta | x)$ denotes the probability of all of your observed data, not just one value. Also the "$|$" may cause confusion as it ...


6

A simple sanity-check on whether an image classifier can perform a task in theory is: Can a human expert, using the same image plus a list of catgeories that they are familiar with, perform the same task? It is important you only consider the contents of the image (or in general the data you are prepared to supply to the classifier) and the expert's ...


5

The human brain works by having neurons constantly fire at different rates. So, if the firing rate increases, the neuron is transmitting overly exciting or calming information to further neurons connected to it. How other neurons connected to the former neuron respond on the messages sent by it, depends on the strength of the connection between the connected ...


5

It is our "current" target. We assume that the value we get now is at least a closer approximation to the "true" target. We're not so much moving towards a wrong value as we are moving away from a more wrong value. Of course, it is all base on random trials, so saying anything definite (such as: "we are guaranteed to improve at each ...


5

Parametric Methods A parametric approach (Regression, Linear Support Vector Machines) has a fixed number of parameters and it makes a lot of assumptions about the data. This is because they are used for known data distributions. i.e, it makes a lot of presumptions about the data Non-Parametric Methods A non-parametric approach (k-Nearest Neighbours, Decision ...


5

It really depends on what words you put after "ground truth". Sometimes people will talk about "ground truth labels", for example in the context of classification or regression problems. The "ground truth labels" in such a case would refer to the true labels of instances; the labels that we use as target labels for instances ...


5

A model as a set of functions In some cases in machine learning, a model can be thought of as a set of functions, so here's the first difference. For example, a neural network with an arbitrary vector of parameters $\theta \in \mathbb{R}^m$ is often denoted as a model, then a specific combination of these parameters represents a specific function. More ...


4

Generally, if one googles "quantum machine learning" or anything similar the general gist of the results is that quantum computing will greatly speed up the learning process of our "classical" machine learning algorithms. This is correct. A lot of machine learning methods involve linear algebra, and it often takes far fewer quantum ...


4

Normally when you write a program, you are acting like a boss that micromanages the job, telling the workers how to accomplish a task, perhaps without even letting them know what the purpose is. What you are hoping to be is a boss that gives the workers a goal and allows them to determine how to accomplish it. In many ways, that is one of the aims of AI. We ...


4

Let's quickly get out our copies of Deep Learning by Goodfellow et al. (2016). More specifically, I'm referring to page 276. On this page, the authors argue for a relatively small minibatch size, since there are less than linear returns for estimating the gradient when increasing the minibatch size. Returns here refer to the reduction of the standard error ...


4

As it is referred in the survey paper "Active Learning Literature Survey": The key idea behind active learning is that a machine learning algorithm can achieve greater accuracy with fewer training labels if it is allowed to choose the data from which it learns. An active learner may pose queries, usually in the form of unlabeled data instances to ...


4

If the learning rate is greater than or equal to $1$ the Robbins-Monro condition $$\sum _{{t=0}}^{{\infty }}a_{t}^{2}<\infty\label{1}\tag{1},$$ where $a_t$ is the learning rate at iteration $t$, does not hold (given that a number bigger than $1$ squared becomes a bigger number), so stochastic gradient descent is not generally guaranteed to converge to a ...


4

In machine learning, the term bias can refer to at least 2 related concepts A (learnable) parameter of a model, such as a linear regression model, which allows you to learn a shifted function. For example, in the case of a linear regression model $y = f(x) = mx + b$, the bias $b$ allows you to shift the straight-line up an down: without the bias, you would ...


4

It's the same thing, first version is the special case of the more general one. In the first case you only have two classes, it's binary cross-entropy, and they also included iteration over batch of samples. In the second case you have multiple classes and in the current form it's only for a single sample. In the first case there is only one output, if you ...


4

They are equivalent. When we consider a particular instance as a vector, we are not literally imagining it as an arrow with it's head at the point coordinates and tail at the origin. It's just when you are working with a tuple of numbers in a mathematical context, it is conventional to call it a vector. This language follows into machine learning which is ...


4

Let me address first some of the things you wrote in your question: There are certain proteins that contain metal components, known as metalloproteins. Natural proteins do not ever contain metal components as far as we know. Natural proteins are composed of natural amino acids which only contain H,C,O,N,S. Selenocysteine contains Se (also a non-metal!) but ...


4

There are a few possible approaches to deploying a ML model to a microcontroller. The main limiting factor to deployment on microcontollers is that ML models are usually a representation of a set of parameters that are intended to be used as input to a prediction algorithm alongside a new datapoint. Most such models assume the presence of an accompanying ...


4

You can already do this with some neural networks, such as GANs and VAEs, which are generative models that learn a probability distribution over the inputs, so they learn how to produce e.g. images that are similar to the images they were trained with. Now, if you're interested in whether there is a black-box method, i.e. a method that, for every possible ...


4

In terms of transfer learning, semantic gap means different meanings and purposes behind the same syntax between two or more domains. For example, suppose that we have a deep learning application to detect and label a sequence of actions/words $a_1, a_2, \ldots, a_n$ in a video/text as a "greeting" in a society A. However, this knowledge in Society ...


3

CNNs learn convolutional filters that get trained on finding local, recurring patterns in some kind of image/volume data. 1D convolution is actually a thing, but I think what would be more suitable for your case is using Recurrent Neural Nets. They are specifically designed for working on time series-es of heterogeneous data. Update: I would like to ...


3

In a classification problem it's better to get higher error and higher error slope when we predict the label wrong. As you see in the graph by using cross-entropy you get high error when the algorithm predict a label wrong and small error when the prediacted label is close enough, so it helps us to separate the predicted classes better.


3

There's some useful information in your description, but that's just a very vague description of how neural networks with sigmoid activation functions are trained. Moreover, there are many other AI systems apart from neural networks (such as support vector machines, expert systems, etc.), which, of course, I cannot exhaustively list here. Is my ...


3

In statistics, if $X$ and $Y$ are independent and randomly distributed variables: $\mathbb{E}[X + Y] = \mathbb{E}[X] + \mathbb{E}[Y] \\ Var(X + Y) = Var(X) + Var(Y) \\ \mathbb{E}[XY] = \mathbb{E}[X]\mathbb{E}[Y] \\ Var(XY) = (Var(X) + \mathbb{E}[X]^2)(Var(Y) + \mathbb{E}[Y]^2) - \mathbb{E}[X]^2\mathbb{E}[Y]^2$ Let $Q$ and $K$ be random $d_k$ x $d_k$ matrices,...


3

Can deep reinforcement learning algorithms be deterministic in their reproducibility in results? Yes, but only if you control all places in the code where stochastic methods are used (typically by seeding the affected RNGs): Neural network weight initialisation Action choice for $\epsilon$-greedy or other behaviour policy (does not apply in your case, ...


3

A first question that I think is important to consider is: do you expect the data that you're dealing with to be changing over time (i.e. do you expect there to be concept drift)? This could be any kind of change. Simply changes in how frequent certain inputs are, changes in how frequent positives/negatives are, or even changes in relations between inputs ...


Only top voted, non community-wiki answers of a minimum length are eligible