54

The black box thing has nothing to do with the level of expertise of the audience (as long as the audience is human), but with the explainability of the function modelled by the machine learning algorithm. In logistic regression, there is a very simple relationship between inputs and outputs. You can sometimes understand why a certain sample was incorrectly ...


21

A baseline is the result of a very basic model/solution. You generally create a baseline and then try to make more complex solutions in order to get a better result. If you achieve a better score than the baseline, it is good.


19

While I agree on ncasas answer in most points (+1), I beg to differ on some: Decision Trees can be used as black box models, too. In fact, I'd say in most cases they are used as black-box models. If you have 10,000 features and a tree of depth of 50 you cannot reasonably expect a human to understand it. Neural Networks can be understood. There are many ...


11

Howard Dresner, in 1989, is believed to have coined the term "business intelligence", to describe "concepts and methods to improve business decision making by using fact-based support systems.". When he was at Gartner Group. This is a common mantra, spread over the Web. I have not been able to trace the exact source for this origin yet. Many insist on he was ...


11

A baseline is a method that uses heuristics, simple summary statistics, randomness, or machine learning to create predictions for a dataset. You can use these predictions to measure the baseline's performance (e.g., accuracy)-- this metric will then become what you compare any other machine learning algorithm against. In more detail: A machine learning ...


8

In Sepp Hochreiter's original paper on the LSTM where he introduces the algorithm and method to the scientific community, he explains that the long term memory refers to the learned weights and the short term memory refers to the gated cell state values that change with each step through time t. edit: quote from paper "Recurrent networks can in principle ...


8

(I actually wanted to write this as an answer to the Cross Validated question: Difference between Anomaly and Outlier, but the question is protected - I think answering it here should be fine, despite the lower visibility) People occasionally argue that there is no difference between an outlier and an anomaly by citing Charu Aggarwal, author of the Book "...


8

Polynomial regression (for nth degree polynomial) in statistics is a special case of linear regression. Lets give an example for square function: 1. y = w*x This is linear in terms of both weight (w) and data (x). 2. y = w*(x^2) OR y = w*z ; where z = x^2 This is still linear in terms of weight (w) and still treated as a ...


8

I found this article by Cynthia Rudin which goes a bit more into the difference between the two terms that is in line with your source from O'Rourke. At the core it is about the time and mechanism of the explanation: A priori (Interpretable) vs. A posterio (Explainable) I found this quote to be very helpful and inline with my own thoughts (emphasis mine): ...


7

If you assume duck to mean "irrelevant decorative elements" there are a few things that strike me as likely: (1) the "squares" only roughly indicate location of the target area; (2) squares represent volume, which is incongruent when overlaid on 2d geography; (3) shading of mountains/geographic features doesn't add detail. Also problematic: comparison of ...


7

"Raw data" is what we say in NLP.


7

According to tensorflow documentation about CNN, The first abstraction we require is a function for computing inference and gradients for a single model replica. In the code we term this abstraction a "tower". To get the relevant context and more, check this.


7

It comes down to model interpretability and explainability. Given the output of a simpler model, it is possible to identify exactly how each input contributes to model output, but that gets more difficult as models get more complex. For example with regression you can point to the coefficients, with a decision tree you can identify the splits. And with this ...


7

Fundamentally there is no difference. Say you have data and you want to build a model of it. As the name suggests, modeling is about finding a model, that is, a simplified representation of your data. In turn, we can view the model as an underlying process that generated your data in the first place, plus some noise. From that point of view, the data you ...


6

The error function is the function representing the difference between the values computed by your model and the real values. In the optimization field often they speak about two phases: a training phase in which the model is set, and a test phase in which the model tests its behaviour against the real values of output. In the training phase the error is ...


6

There are three terms from social science that apply to your situation: Reflexivity - refers to circular relationships between cause and effect. In particular, you could use the definition of the term adopted by George Soros to refer to reverse causal loop between share prices (i.e. present value of fundamentals) and business fundamentals. In a way, the ...


6

If you look at older machine learning algorithms, they rely on the input being a feature and learn a classifier, regressor, etc on top of that. Most of these features were hand crafted, meaning, they were designed by humans. Classical examples of features in computer vision include Harris, SIFT, LBP, etc . The problem with these is that they were designed ...


6

Is the "expected reward" actually $\mathcal{R}^a_{ss'}$ instead of $V^\pi(s)$? In short, yes. Although there is some context associated - $\mathcal{R}^a_{ss'}$ is in the context of specific action and state transition. You will also find $\mathcal{R}^a_{s}$ used for expected reward given only current state and action (which works fine, but moves around ...


6

I think it is helpful to distinguish between linear functions (representing the relationship between independent and dependent variables) and linear models (representing the relationship between the model parameters and the outcome). A linear model can be represented by a non-linear function. A linear regression model is any model that is represented by a ...


5

Effectively, Word2Vec/Doc2Vec is based on distributional hypothesis where the context for each word is its nearby words. Similarly, LSA takes the entire document as the context. Both techniques solve the word embedding problem - embed words into a continuous vector space while keeping semantically related words close together. On the other hand, LDA isn't ...


5

I think the term "sensitivity" comes from the world of medical tests. A very sensitive test will test positive for most or all people who take the test and really have a disease, as well as for many people who don't. This corresponds to high recall, which means the query retrieves most or all of the relevant documents, as well as many that may not be ...


5

Black box models refer to any mathematical models whose equations are chosen to be as general and flexible as possible without relying on any physical/scientific laws. Grey box models are mathematical models where part of the equations (mathematical function) comes from physical known laws but the remaining part is assumed general function to compensate for ...


5

I will try to answer this question conceptually and not technically so you get a grasp of the mechanisms in RL. Bootstrapping: When you estimate something based on another estimation. In the case of Q-learning for example this is what is happening when you modify your current reward estimation $r_t$ by adding the correction term $\max_a' Q(s',a')$ which is ...


5

Yes! It seems like first statistics reference to over fitting appears in The Quarterly Review of Biology . It says "Perhaps we are old fashioned but to us a six-variate analysis based on thirteen observations seems rather like overfitting. I am attaching the screenshot of that particular page for your reference.


4

It is a bad idea to counterpose "unstructure data" to, say, tabular data (as in "non-tabular data"), as you will have to elliminate other alternatives as well (e.g., "non-tabular and non-graph and ... data"). "Plain text" (-- my choice) or "raw text" or "raw data" sound fine.


4

ML engineers don't know what goes on inside a neural net Sorry to contradict you, but it's true. They know how neural networks learn, but they do not know what any given neural network has learned. The logic learned by neural networks is notoriously inscrutable. The point of using machine learning is usually to learn the rules that a programmer or domain ...


4

When the data is linearly inseparable, we use MLP. Here what is meant by "data"--is it the response or the input feature that is linearly inseparable? This means that a linear function of the input features is unable to separate the response. To answer your question a bit more directly: Given only a linear function of the inputs, the response is the thing ...


3

Recall means to bring back or remember. The terminology comes from information retrieval where it's usually being applied to a result set from a query. I suppose the sense of it is, how much of the set of right answers was retrieved by the query? how much of it was recalled? I don't know if "coverage" is better or not. The word "sensitivity" is also used ...


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