I’m not necessarily good at it, but most days I get to solve the puzzle.
The experience is completely different with Semantle — a Wordle-inspired puzzle in which you also need to guess the word of the day.
Unlike in Wordle, Semantle gives you unlimited guesses though. And, boy, you will need many!
Like Wordle, Semantle gives you hints as to how close your guesses were to the secret word of the day.
However, where Wordle shows you how good your guesses were in terms of the letters used, Semantle evaluates the semantic similarity of your guesses to the secret word. For the 1000 most similar words to the secret word, it will show you its closeness like in the picture above.
This semantic similarity comes from the domain of Natural Language Processing — NLP — and this basically reflects how often words are used in similar contexts in natural language.
For instance, the words “love” and “hate” may seem like opposites, but they will often score similarly in grammatical sentences. According to the semantle FAQ the actual opposite of “love” is probably something like “Arizona Diamondbacks”, or “carburetor”.
Another example is last day’s solution (15 March 2022), when the secret word was circle. The ten closest words you could have guessed include circles and semicircle, but more distinctive words such as corner and clockwise.
Further downfield you could have guessed relatively close words like saucer, dot, parabola, but I would not have expected words like outwaited, weaved, and zipped.
The creator of Semantle scored the semantic similarity for almost all words used in the English language, by training a so-called word2vec model based on a very large dataset of news articles (GoogleNews-vectors-negative300.bin from late 2021).
Now, every day, one word is randomly selected as the secret word, and you can try to guess which one it is. I usually give up after 300 to 400 guesses, but my record was 76 guesses for uncovering the secret word world.
Gordon finds that there are four main features of the R programming language that are essential to his work and in a sense unique to the R language. Here they are, along with quotes by Gordon explaining R’s unique selling points in his words:
(1) Native data science structures
It’s relatively easy to do data science in R without any external libraries. You can read data from a csv into a data frame, plot and clean that data, and analyse it using built-in statistical models.
(2) Non-standard evaluation
Non-standard evaluation lets you do things like use a variable name in a plot title, or evaluate a user-supplied expression in a different environment.
For example, R lets you specify models with a formula interface like this: lm(mtcars, mpg ~ cyl). This is a natural way for statisticians to specify statistical models because they’re usually familliar with the syntax, but without NSE there’s no way to make that function work as written because mpg and cylare not objects in the calling environment.
(3) Packaging concensus
R let me get up and running, installing packages, filtering data, and printing plots in under 20 minutes, which meant that I stayed interested in the language and eventually started using it professionally. I had actually started to learn Python at around the same time but just found it too difficult. […]
The user that I care the most about only has 20 minutes of attention and no real programming skill, so the only thing they can “just” do is copy and paste one line of code into a console. If that doesn’t work, I’ve lost them, and they’ll spend another lonely year renewing their SPSS licenses.
(4) Functional programming
I really like this pattern of [functional] programming because breaking complicated jobs down into small functional bricks gives me confidence that the overall solution is correct. I can work on the small functions, verify that they’re correct through tests, and then know that combining those building blocks together won’t change their behaviour.
Although I personally do not fully agree with these four points (e.g., I very much like to leverage functional programming in Python and it works like a charm!) I very much liked the outline Gordon provides. I’d love to hear your thoughts as well, so do share them in the comments.
For now, let’s end with some other lovely quotes by Gordon:
The thing is, I don’t use R out of some blind brand loyalty but because I don’t like working hard.
I came to R from an Excel background, and for a long time I had internalized the feeling that serious engineers used Python, while analysts or researchers could use languages like R. Over time I’ve realized that the people making that statement often aren’t really informed. They rarely know anything about R, and often don’t really write production-quality code themselves.
In contrast, most of the very senior engineers I’ve met understand that all programming languages are basically just bundles of trade-offs, and so no single language is going to be globally superior to another. There really are no production languages – only production engineers.
If you have ever programmed in R, you are probably familiar with the Grammar of Graphics due to ggplot2. You can read more about the Grammar of Graphics here, but the general idea behind it is that visualizations can be build up through various layers, each of which have certain characteristics (aesthetics in ggplot2).
This post is not about ggplot2 nor specifically the Grammar of Graphics, but rather a summary of the official release of Vega-Lite 2, a high-level language for rapidly creating interactive visualizations which you might know from the R-package ggvis.
Vega-Lite has four operators to compose charts: layer, facet, concat and repeat. Layer stacks charts on top of each other in an orderly fashion. Facet divides and charts the data into groups. Concat combines multiple charts into dashboard layouts and, finally, repeat concatenate charts. Most importantly is that these operators can be combined! The example below compares weather data in New York and Seattle, layering data for individual years and averages within a repeated template for different measurements.
Vega-Lite 2 is especially useful because of the included interaction options. Programmers can specify how users can interactive select the data in their visualizations (e.g., a point or interval selection), along with possible transformations. With these interactions, users can for instance filter data, highlight points, or pan or zoom a plot. The plot below uses an interval selection, which causes the chart to include an interactive brush (shown in grey). The brush selection parameterizes the red guideline, which visualizes the average value within the selected interval.
However, this is not all! When multiple visualizations are combined in a dashboard, interactive selections can apply to all. Below, you see an interval selection being applied over a set of histograms. As a viewer adjusts the selection, they can immediately see how the other distributions change in response.
This blog summarized work that has been posted here, here, and here.
Iain of degeneratestate.org wrote a three-piece series where he applied text mining to the lyrics of 222,623 songs from 7,364 heavy metal bands spread over 22,314 albums that he scraped from darklyrics.com. He applied a broad range of different analyses in Python, the code of which you can find here on Github.
For example, he starts part 1 by calculated the difficulty/complexity of the lyrics of each band using the Simple Measure of Gobbledygook or SMOG and contrasted this to the number of swearwords used, finding a nice correlation.
Furthermore, he ran some word importance analysis, looking at word frequencies, log-likelihood ratios, and TF-IDF scores. This allowed him to contrast the word usage of the different bands, finding, for instance, one heavy metal band that was characterized by the words “oh yeah baby got love“: fans might recognize either Motorhead, Machinehead, or Diamondhead.
Using cosine distance measures, Iain could compare the word vectors of the different bands, ultimately recognizing band similarity, and song representativeness for a band. This allowed interesting analysis, such as a clustering of the various bands:
However, all his analysis worked out nicely. While he also applied t-SNE to visualize band similarity in a two-dimensional space, the solution was uninformative due to low variance in the data.
He could predict the band behind a song by training a one-vs-rest logistic regression classifier based on the reduced lyric space of 150 dimensions after latent semantic analysis. Despite classifying a song to one of 120 different bands, the classifier had a precision and recall both around 0.3, with negligible hyper parameter tuning. He used the classification errors to examine which bands get confused with each other, and visualized this using two network graphs.
His first approach was to use probabilistic distributions known as language models. Basically he develops a Markov Chain, in his opinion more of a “unsmoothed maximum-likelihood language model“, which determines the next most probable word based on the previous word(s). This model is based on observed word chains, for instance, those in the first two lines to Iron Maiden’s Number of the Beast:
Another approach would be to train a neural network. Iain used Keras, which ran on an amazon GPU instance. He recognizes the power of neural nets, but says they also come at a cost:
“The maximum likelihood models we saw before took twenty minutes to code from scratch. Even using powerful libraries, it took me a while to understand NNs well enough to use. On top of this, training the models here took days of computer time, plus more of my human time tweeking hyper parameters to get the models to converge. I lack the temporal, financial and computational resources to fully explore the hyperparameter space of these models, so the results presented here should be considered suboptimal.” – Iain
He started out with feed forward networks on a character level. His best try consisted of two feed forward layers of 512 units, followed by a softmax output, with layer normalisation, dropout and tanh activations, which he trained for 20 epochs to minimise the mean cross-entropy. Although it quickly beat the maximum likelihood Markov model, its longer outputs did not look like genuine heavy metal songs.
So he turned to recurrent neural network (RNN). The RNN Iain used contains two LSTM layers of 512 units each, followed by a fully connected softmax layer. He unrolled the sequence for 32 characters and trained the model by predicting the next 32 characters, given their immediately preceding characters, while minimizing the mean cross-entropy:
“To generate text from the RNN model, we step character-by-character through a sequence. At each step, we feed the current symbol into the model, and the model returns a probability distribution over the next character. We then sample from this distribution to get the next character in the sequence and this character goes on to become the next input to the model. The first character fed into the model at the beginning of generation is always a special start-of-sequence character.” – Iain
This approach worked quite well, and you can compare and contrast it with the earlier models here. If you’d just like to generate some lyrics, the models are hosted online at deepmetal.io.
In part 3, Iain looks into emotional arcs, examining the happiness and metalness of words and lyrics.
When applied to the combined lyrics of albums, you could examine how bands developed their signature sound over time. For example, the lyrics of Metallica’s first few albums seem to be quite heavy metal and unhappy, before moving to a happier place. The Black album is almost sentiment-neutral, but after that they became ever more darker and more metal, moving back to the style to their first few albums. He applied the same analysis on the text of the Harry Potter books, of which especially the first and last appear especially metal.
The first programs for (scientific) text mining are already over 50 years old. More recent efforts, such as the Linguistic Inquiry Word Count (LIWC; Tausczik & Pennebaker, 2010), have greatly improved our text analytical capabilities. Moreover, several single-purpose programs have been developed, which also consider syntactic text structures (e.g., Syntactic Complexity Analyzer [Lu, 2010], TAALES [Kyle & Crossley, 2015]).However, the widespread use of many of these programs has been hampered by two major barriers.
First, considerable technical expertise is required, which obstructs researchers without statistical backgrounds. For example, packages such as tm in R (Meyer et al., 2015) have been developed to conduct natural-language processing, but the steep learning curve forms a challenge. Additionally, the constant increase of computational processing power and the proliferation of new algorithms makes it difficult for researchers to maintain working knowledge of state-of-the-art methods.
Alternatively, most of the existing user-friendly NLP programs (and packages), such as RapidMiner (Akthar & Hahne, 2012), SAS Text Miner (Abell, 2014), or SPSS Modeler (IBM Corp., 2011), charge either a large software fee up front or a subscription fee. The cost of these programs can be prohibitively expensive for junior researchers and researchers looking to integrate new techniques into their research toolbox.
In the attached article, TACIT is introduced: Text Analysis, Crawling and Investigation Tool. TACIT is an open-source architecture that establishes a pipeline between the various stages of text-based research by integrating tools for text mining, data cleaning, and analysis under a single user-friendly architecture. In addition to being prepackaged with a range of easily applied, cutting-edge methods, TACIT’s design also allows other researchers to write their own plugins.
The authors’ hope is that TACIT can facilitate the integration and use of advancements in computational linguistics in psychological research, and by doing so can help researchers make use of the ever-growing documents of our social discourse in ways that have previously not been possible.