Is this you? As a journalist, you can create a free Muck Rack account to customize your profile, list your contact preferences, and upload a portfolio of your best work.
Claim your profile
Articles by Arthur Mello
LLMs Don’t Think… Right?
Press enter or click to view image in full size Rethinking what it means to think in the age of AI We’ve all heard it: “LLMs don’t think. They just predict the next token.” That second part is an oversimplification, but that doesn't really matter here. Let’s talk about the first part. Here’s the case for why LLMs do think , through three angles: Definition Process Output Definition To say “LLMs don’t think,” we need to define what thinking even means.
Bolsonaristas provam do próprio veneno
O motim bolsonarista em defesa da pauta da anistia na Câmara extrapolou os limites determinados no Regimento Interno (PDF – 2 MB) da Casa, o que resultou em pedidos de sanção contra os congressistas envolvidos. Estratégias para obstrução de pautas são ações regimentais determinadas no Regimento. O que vimos, no entanto, foram atos antidemocráticos e de chantagem contra o presidente da Câmara.
Do Labor Regulations Slow Down Innovation?
A paper by Aghion, Bergeaud & Van Reenen explores how firm-size-based labor laws affect innovation in France. Spoiler: regulations discourage incremental innovation — but encourage radical bets. Key finding: Badly designed labor regulation reduces aggregate innovation by 5.4%. In France, firms crossing 50 employees must comply with stricter rules: works councils, union representation, health & safety committees, and more. These burdens act like a “regulatory tax” on growth.
Berkson’s Paradox: Why Your Data Might Be Lying to You
Have you ever felt like there’s a strange trade-off in the people you date — the kinder they are, the less attractive they seem? Or maybe doctors notice that patients either have diabetes or high blood pressure, but rarely both? That’s Berkson’s Paradox in action. It’s a statistical illusion that happens when we only look at a filtered or biased sample. The relationships we see inside that sample can be completely different (even reversed) from what’s true in the broader population.
Opinião - Carolina Ricardo, Julia Neiva e Arthur Mello: Por uma quarentena maior para os policiais candidatos
Nas eleições de 2024, o Policial Machado, candidato a vereador em São Paulo, divulgou vídeo em que um tenente ligado ao 8º Batalhão de Polícia Militar apoiou explicitamente sua candidatura. Machado mencionava também o apoio do batalhão, que seria uma espécie de "QG da sua campanha". Casos como esse representam um flagrante uso irregular das instituições policiais em favor de uma campanha política.
AI Is Not Taking Our Jobs
Image created by author, using MidJourney Member-only story A quantitative exploration of the relationship between technology, labor and wealth Arthur Mello · Follow Published in Data Science Collective · 9 min read · -- I see many people say AI will take our jobs any time now, based on the following narrative: the better AI becomes, the less companies will need us, therefore we will be replaced. It’s an interesting chain of thought, but does it have the empirical basis to back it up?
How to Become an AI Expert in 5 minutes
Arthur Mello · Follow Published in The Startup · 2 min read · -- Do you want to impress colleagues, dazzle recruiters, or dominate a panel discussion without actually knowing anything about AI? Good news! In just five minutes, you too can sound like an AI expert. Forget years of research, hands-on coding, or grueling math — this guide will teach you everything you need to fake it until you make it in the world of AI. Say “garbage in, garbage out.” It means nothing, but it sounds cool.
Opinião - Flávia Pellegrino e Arthur Mello: Não aceitem o autoritarismo
Os ventos de uma democracia forte sopram com direção e sentidos bem definidos: o respeito ao processo eleitoral e a responsabilização daqueles que tentaram abolir o Estado democrático de Direito. Recentemente, a Folha de S.Paulo tem servido de palco a um processo que vem acontecendo há algum tempo de normalização de lideranças autoritárias e de esquecimento das conspirações antidemocráticas e golpistas que vivemos nos últimos anos.
The Power of Creative Destruction
How innovation drives economic growth With all the buzz surrounding AI and its impact on jobs, understanding the forces that shape our economy is more important than ever. Elected “Best Book of the Year” by The Economist, The Power of Creative Destruction revisits the Schumpeterian framework, putting innovation at the heart of economic growth and exploring how it disrupts and transforms industries. Traditional economic models, like the Solow model, view growth as a result of capital accumulation.
Why You’re Struggling to Break into Data Science
And what you can do about it I see many people having a hard time transitioning from other fields into Data Science, even though there’s more open jobs every day. Many companies end up going for people who already have domain experience or who are fresh out of college, with a degree in Statistics or Computer Science. Although it can be hard to make this transition, I did it, and I think you can do it too, as long as you have the right strategy.
Opinião | As eleições e a valiosa oportunidade de defesa do SUS
Com a chegada das eleições municipais de 2024, é crucial refletir sobre os desafios da saúde pública no Brasil. A atual situação exige um debate aprofundado sobre as políticas e práticas necessárias para assegurar um sistema de saúde mais justo e acessível a todos. Nos últimos anos, o Brasil tem enfrentado crises significativas na área da saúde.
Marginal Structural Models
Causal inference for time-varying treatments Marginal Structural Models (MSMs) are particularly useful in the analysis of longitudinal data with time-varying treatments. They are designed to address situations where the effects of treatments or interventions change over time, and standard regression models fail to account for these dynamic effects due to their inability to handle time-dependent confounding.
Methodological quality of research on perioperative immunomodulatory supplementation in oncological gastrointestinal tract surgery: a meta-research protocol
Introduction Research in nutrition is among the most controversial areas of science and has gained prominence in meta-research studies, a relatively new field that aims to analyse scientific questions that are not yet fully understood, aspects of the scientific method that need improvement, or those that have not been assessed with appropriate methodology.1 2 Among the nutrition topics that exhibit divergence of opinions is the recommendation of an immunomodulatory diet with arginine,...
Survival Analysis in Python: A Step-by-Step Guide for Churn Prediction
Improve customer retention with time-to-event data Survival analysis is a branch of statistics used to model and predict the time until an event occurs, commonly used in medicine to understand the factors that impact someone’s lifespan or the time until a disease progresses. It focuses on time-to-event data, where the exact time of an event’s occurrence is known, unlike traditional classification techniques that typically deal with binary dependent variables.
Books for 2024
A few of my 2023 reads, and why I recommend them for your 2024 reading list :) “Everybody Writes” by Ann Handley is a refreshing and practical guide that transcends the conventional boundaries of writing advice. Unlike typical writing manuals, Handley not only demystifies the art of writing but also illuminates its crucial role in our digital age.
Variation observed in consensus judgements between pairs of reviewers when assessing the risk of bias due to missing evidence in a sample of published meta-analyses of nutrition research
• None There was a high raw percentage agreement between pairs for most items in the ROB-ME (Risk Of Bias due to Missing Evidence) tool when applied to 42 published meta-analyses of nutrition research, but these estimates had wide confidence intervals. Furthermore, Gwet’s Agreement Coefficient (ACs), which are corrected for chance, indicated substantial variation in assessments.
Credibility at stake: only two-thirds of randomized trials of nutrition interventions are registered and lack transparency in outcome and treatment effect definitions
Cookie Preference Center We use cookies which are necessary to make our site work. We may also use additional cookies to analyse, improve and personalise our content and your digital experience. For more information, see our Cookie Policy and the list of Google Ad-Tech Vendors. You may choose not to allow some types of cookies. However, blocking some types may impact your experience of our site and the services we are able to offer.
The Ultimate Workflow
Are you tired of starting from scratch every time you need to explore your data, without a clear roadmap? Look no further! I will guide you through a step-by-step process using Python to uncover valuable insights and trends hidden in your data. Whether you're a beginner or an experienced data analyst, this article has something for you. We will look at different ways in which you can explore your dataset, depending on your end goal, with clear steps and objectives for each of them.
Factor Analysis for Marketing with Python
Factor analysis is a statistical technique used to identify the main factors that affect a particular outcome. It involves breaking down a complex phenomenon into its individual components, and then analyzing the effect of each component on the outcome. Take the image below as an example: In this case, we have three observed variables: "Average spending", "Customer satisfaction" and "Web site visits", and two latent factors: "Price perception" and "Product perception".
Data Analysis for Marketing: an Overview
By Arthur Mello Uncovering insights and patterns through data analysis Exploratory data analysis is a crucial step in the data science workflow, allowing us to understand, summarize, and visualize our data in order to uncover patterns and relationships. By using a variety of statistical and graphical methods, we can gain valuable insights that can...
Data Science for Business
Data Science for Business, published in 2013, aims to introduce data science from a business perspective, achieving a good balance between broadness and technical detail. This was possible thanks to the successful combination of two authors: a data scientist - Tom Fawcett holds a PhD in Machine Learning and over twenty years of experience in research and development - and a business professor - Foster Provost teaches business analytics and data science for MBA students at NYU.
Should You Take the Google Data Analytics Certification?
Whether you are looking for an entry-level job in the analytics field, or just trying to understand how Google does data analysis, this certification might be the right one for you! It includes 8 courses, and the estimated time to complete it is approximately 6 months, considering a pace of 10h per week (although I think that's an overestimation).
Causal Inference: an Overview
Find out when correlation actually means causation First of all, why should we bother studying causation? Isn’t correlation enough for making predictions? Hasn’t machine learning made the study of causation pointless? Well, not really. Many times, we are interested in a relationship of this kind: “if I do X, then Y will happen”, which is not the same as “if X happens, then Y will happen”.
Regularization for Machine Learning
Why it’s one of the most important techniques, and how to use it Regularization is a type of technique that calibrates machine learning models by making the loss function take into account feature importance. Intuitively, it means that we force our model to give less weight to features that are not as important in predicting the target variable, and more weight to those which are more important.
A New, Better Version of the K-Nearest Neighbors Algorithm
Photo by Nina Strehl on Unsplash Improve your results with KNN by using regularization The other day, while writing this article on the KNN algorithm, I noticed one thing: distances between observations are calculated without taking into account the importance of each feature to the overall task.
7 Data Science Libraries That Will Make Your Life Easier in 2022
Photo by Florian Olivo on Unsplash When doing Data Science, you can end up wasting a lot of time coding and waiting for the computer to run something. I have selected a few Python libraries that can save you time in both situations. Even if you incorporate just one of them to your arsenal, you can still save precious time next time you work on a project. Optuna is an open source hyperparameter optimization framework.
8 Data Science Side Projects for 2022
A “sports betting tipster service” is a service that sells tips on sports bets. People are willing to pay good money if you can regularly provide them with good insights on the next football match. Ok, but what does that have to do with data science? Well, with the right data and the right skills, you are actually in a good position to calculate probabilities regarding many things, including the outcomes of sports events.
Why Machine Learning Engineers are Replacing Data Scientists
By Arthur Mello, Data Science Consultant at AVISIA. Photo by T K on Unsplash. Isn’t ML engineering just another buzzword? ML engineering and data science are not the same things, and here’s why: you know when people say that data science is a mix between business knowledge, statistics, and computer science? Well, ML engineering is a lot more about computer science and less about statistics and business knowledge.
Bayesian Statistics 101
Photo by Jens Lelie on Unsplash Bayesian statistics differs from classical statistics (also known as frequentist) basically in its interpretation of probability. The former sees it as a “degree of belief”, whereas the latter sees it as the “relative frequency observed during many trials”. This difference might seem too abstract, but it has great practical impact on the methods developed afterwards.
Why Machine Learning Engineers are Replacing Data Scientists
Photo by T K on Unsplash And what you should do about it ML engineering and data science are not the same thing, and here’s why: you know when people say that data science is a mix between business knowledge, statistics and computer science? Well, ML engineering is a lot more about computer science, and less about statistics and business knowledge.
The One Data Science Tool You Should Master in 2022
Photo by Dominik Lange on Unsplash Straight to the point: if you don’t know Dataiku DSS, you should. Seriously. It will save you hundrends of hours, multiple headaches and make you enjoy data science even more! DSS (short for Data Science Studio) is an end-to-end machine learning platform created by a company called Dataiku. Sometimes you will see people referencing the platform itself as Dataiku.
K-Nearest Neighbors: Theory and Practice
Photo by Emily-Jo Sutcliffe on Unsplash K-Nearest Neighbors, also known as KNN, is probably one of the most intuitive algorithms there is, and it works for both classification and regression tasks. Since it is so easy to understand, it is a good baseline against which to compare other algorithms, specially these days, when interpretability is becoming more and more important. Say you want to sell your house, and need to estimate a price for it.
How I Passed the GCP Professional ML Engineer Certification
Photo by Billy Huynh on Unsplash In 2020, Google Cloud Platform released its latest certification: the Professional ML Engineer Certification. When 2021 started, I decided to give it a try, even though Google recommends 3+ years of practical experience before taking the test, and I had actually never worked with GCP in real life. But why do it, then?
4 Habits of Bad Data Analysts
1. Neglects Data The worst bad habit I’ve seen is data analysts that don’t do any exploratory data analysis ( EDA ) on new tables and jump immediately to writing SQL. When I was a consultant I had a new colleague join two tables together without first looking at the data to realize there were duplicates on the join key. This caused a cartesian join that ran for hours and halted the daily ETL jobs until the DBA killed the query.
Feature Engineering Examples: Binning Categorical Features
Working with categorical data for machine learning (ML) purposes can sometimes present tricky issues. Ultimately these features need to be numerically encoded in some way so that an ML algorithm can actually work with them. You’ll also want to consider additional methods for getting your categorical features ready for modeling. For example, your model performance may benefit from binning categorical features. This essentially means lumping multiple categories together into a single category.
Hypothesis Testing
Photo by ThisisEngineering RAEng on Unsplash A hypothesis is a claim or a premise that we want to test. Hypothesis testing is a way of backing upyour conclusions with data, in a more “scientific” way. It is useful not only to scientists, but actually important in so many areas, ranging from marketing to web design to pharmaceutical trials and much more.
Sampling Methods for Data Science
Photo by Megan McClain on Unsplash What is the best way to get samples out of a population? In most studies, it is pretty hard (or sometimes impossible) to analyse a whole population, so researchers use samples instead. In statistics, survey sampling is the process by which we get a sample from our population, in order to conduct a survey. As data scientists, we usually use data that was previously collected, so we don’t spend too much time thinking about how to actually do this.
The Gospel of Misanthropy
Microaggression, Privilege Interrogation and White Fragility are devices of a emergent cult whose target is the human psyche Misanthropy comes in many forms. Some who suffer abuse want revenge against the world. For others, this animosity is channeled into an idiosyncratic hatred of specific targets such as women or gays. For still others misanthropy is born out of dissonance between an idealized notion of humanity and observed reality.
Can We Make Artificial Intelligence More Ethical?
Photo by Andy Kelly on Unsplash How to tackle the most pressing issues when it comes to ethics in AI and robotics What are the most pressing issues when it comes to ethics in AI and robotics? How will they affect the way we live (and work)? Sooner or later these issues will concern you, whether you work in the field or not. Here we will go through the main ideas contained in the paper Robot ethics: Mapping the issues for a mechanized world, while I add some of my own input.
Craig Newmark Philanthropies Donates to NYU’s Center for Social Media and Politics (CSMaP) and NYU’s Center for Data Science
We’re thrilled to announce that CDS and NYU’s Center for Social Media and Politics have received $400,000 from Craig Newmark Philanthropies, an organization that supports public charities and other organizations in areas such as journalism, voter protection, women in technology, veterans and military families. $350,000 is going to CDS student support, and $50,000 is going towards Politics student support. This donation will fund our PhD students and enable them to continue their research projects.
Creativity and Artificial Intelligence
Photo by Franck V. on Unsplash How does AI generate new ideas? Will it ever replace human creativity? Artificial intelligence is one of those buzzwords that we see everywhere now. It can be used to detect credit card fraud, play chess and even drive cars. But these are all tasks which we are more or less willing to give up and admit that computers can perform better than us — it seems reasonable that one day or the other, computers will be driving around, coordinating the traffic in our cities.
Design Thinking for Artificial Intelligence Projects
Photo by Daria Nepriakhina on Unsplash How IBM adapted design thinking principles to build a workflow for AI projects IBM uses an interesting workflow for AI projects that is based on design thinking principles, shown in detail in its specialisation on Coursera. It’s a long MOOC, that goes into thorough detail on implementing AI apps, and its worth taking, if you have the time and technical knowledge.
10 Books Data Scientists Should Read During Lockdown
If you work with data, these books will help you keep active and learning It’s been a few weeks now that the whole world is engaged in fighting the COVID-19 and most people have been locked down for a while. Even though most of us are still working from home, it’s quite easy to get bored during our free time, not sure what to watch next on Netflix. Personally, I love reading, so I thought I’d recommend some books I’ve read lately that might be useful or entertaining for people who work with data.
Data Science and Behavioural Economics
Understanding our biases through data According to Wikipedia, data science “is an inter-disciplinary field that uses scientific methods, processes, algorithms and systems to extract knowledge and insights from structured and unstructured data”. How does it differ from statistics? To be honest, not that much. Statistics is subfield of mathematics that deals with, amongst other things, randomness, uncertainty and information.
Show More
loading
Actions
Is this you?
As a journalist, you can create a free Muck Rack account to customize your profile, list your contact preferences, and upload a portfolio of your best work.Get in touch with Arthur
Contact Arthur, search articles and posts on X, monitor coverage, and track replies from one place.
Learn more about Muck Rack