Blog
How bad is the infiltration of false information?
We are all imperfect messengers. And while I appreciate that many of our more sensible (from my perspective) messages are from messengers who are so imperfect and up in leadership positions with massive followings, I only wish more of the “better” people were more listened to. It seems the useful
Why Federated models in AI leak secrets
I was at a meeting last week discussing how AI models might be used to support archives forefficiency, effectiveness, etc. while maintaining the requirements for those archives to providereliable and authentic records and maintain confidentiality of relevant record data elements. Someone across the room indicated that they were going to
Train-Test split and Cross-validation: Visual Illustrations & Examples
Creating an optimal model that strikes a balance between underfitting and overfitting requires careful consideration. To assess how well our model performs on new, unseen data, we employ the train-test split technique and cross-validation. Train-Test Split: To evaluate a model’s performance, we split the dataset into a training set (approximately
Asking the right questions
I very often see false choices about paths forward. One of the most common ones is the question about the future of humanity regarding AI, robotics, computers, cloning, vaccines, genetic engineering, you name it. Any new area of technology with a tremendous potential is always a multi-edged sword, and someone
Enhancing Accessibility and Inclusivity in Education with ChatGPT
Education is the cornerstone of progress, but for it to be truly impactful, it must be accessible and inclusive to all. In recent years, technology has played a pivotal role in transforming the educational landscape, making strides towards greater accessibility and inclusivity. Among these technological advancements stands ChatGPT, an AI-powered
Ethical Considerations in AI-Powered Education
Did you know that ethical considerations in AI-powered education are becoming increasingly crucial? Or that ensuring data privacy is a top concern in the integration of AI tools in classrooms? It is important for educators to stay on top of ethical considerations in AI-powered education so they can foster responsible
Understanding Confidence Intervals with an Intuitive Example
Confidence intervals (CIs) are a fundamental concept in data science. In this informative guide, we’ll delve into the world of confidence intervals using an intuitive example to help you grasp this concept with confidence. The Bus Stop Scenario: Imagine yourself at a bus stop where the expected arrival time of
Non-linear Relationships: When a 0 Pearson Correlation Coefficient Can Be Surprisingly Meaningful
The Pearson correlation coefficient (denoted as “r”) is widely used in statistics to measure the strength and direction of linear relationships between variables, ranging from -1 (perfect negative linear correlation) to +1 (perfect positive linear correlation). A Pearson correlation coefficient of 0 typically implies the absence of a linear relationship
The do nothing defense
Of course until you are successfully attacked, doing nothing about security is a great approach. It saves money and time and costs you nothing. So lots of folks do nothing, or almost nothing, and I include myself in this latter approach – almost nothing. As little as I can get
Firewalls are alive and well: here’s why
Many years ago, Gartner declared “Firewalls are Dead”, but some 20 years later (give or take) they are still reporting on “next-generation” firewalls. Why is that? Right to the answer Firewalls provide an economy of scale by substantially reducing the input sequences and noise that can contact interfaces. Let’s break
How bad is the infiltration of false information?
We are all imperfect messengers. And while I appreciate that many of our more sensible (from my perspective) messages are from messengers who are so
Why Federated models in AI leak secrets
I was at a meeting last week discussing how AI models might be used to support archives forefficiency, effectiveness, etc. while maintaining the requirements for
Train-Test split and Cross-validation: Visual Illustrations & Examples
Creating an optimal model that strikes a balance between underfitting and overfitting requires careful consideration. To assess how well our model performs on new, unseen
Asking the right questions
I very often see false choices about paths forward. One of the most common ones is the question about the future of humanity regarding AI,
Enhancing Accessibility and Inclusivity in Education with ChatGPT
Education is the cornerstone of progress, but for it to be truly impactful, it must be accessible and inclusive to all. In recent years, technology
Ethical Considerations in AI-Powered Education
Did you know that ethical considerations in AI-powered education are becoming increasingly crucial? Or that ensuring data privacy is a top concern in the integration
Understanding Confidence Intervals with an Intuitive Example
Confidence intervals (CIs) are a fundamental concept in data science. In this informative guide, we’ll delve into the world of confidence intervals using an intuitive
Non-linear Relationships: When a 0 Pearson Correlation Coefficient Can Be Surprisingly Meaningful
The Pearson correlation coefficient (denoted as “r”) is widely used in statistics to measure the strength and direction of linear relationships between variables, ranging from
The do nothing defense
Of course until you are successfully attacked, doing nothing about security is a great approach. It saves money and time and costs you nothing. So
Firewalls are alive and well: here’s why
Many years ago, Gartner declared “Firewalls are Dead”, but some 20 years later (give or take) they are still reporting on “next-generation” firewalls. Why is