“It was not clear, however, how to adapt classical methods of error correction to quantum computers. Quantum information cannot be copied; to correct errors, we need to collect information about them through measurement. The problem is, if you check the qubits, you can collapse their state—that is, you can destroy the quantum information encoded in them. Furthermore, besides having errors in flipped bits, in a quantum computer you also have errors in the phases of the waves describing the states of the qubits.” https://rb.gy/s506p

Should we have AI regulation or should we have a Far West type of internet news where anyone who is fast enough can just shoot whatever he/she likes? Some regulation is important, especially as we create new models which are still very dumb but quite powerful - https://vzocca.substack.com/p/when-a-man-with-a-45-meets-a-man

AI regulation is being discussed both among researchers as well as at the political level. But how did we get here? Whether you believe we are on the cusp of realising truly intelligent machines or we are still in the infancy of intelligent models, whether you believe we already have AGI or whether you think we have Machine Learning based on little more than statistical models, it is important to understand that we have started walking down a path from which we will not be able to come back and that we should start regulating it before it becomes too late, even if we are not there yet - https://vzocca.substack.com/p/a-short-history-of-artificial-intelligence

Suppose you have a thousand-page book, but each page has only a single line of text. You’re supposed to extract the information contained in the book using a scanner, only this particular scanner systematically goes through each and every page, scanning one square inch at a time. It would take you a long time to get through the whole book with that scanner, and most of that time would be wasted scanning empty space.
Such is the life of many an experimental physicist.

In particle experiments, detectors capture and analyze vast amounts of data, even though only a tiny fraction of it contains useful information. “In a photograph of, say, a bird flying in the sky, every pixel can be meaningful,”

But that’s starting to change. With a machine learning tool known as a sparse convolutional neural network (SCNN), researchers can focus on the relevant parts of their data and screen out the rest. Researchers have used these networks to vastly accelerate their ability to do real-time data analysis. https://rb.gy/d4czw

Topologists study the properties of general versions of shapes, called manifolds. Their animating goal is to classify them. In that effort, there are a few key distinctions. What exactly are manifolds, and what notion of sameness do we have in mind when we compare them? https://shorturl.at/DEFY6

Computer scientists can also use “teacher” systems to train another machine to complete a task. But just like with human learning, the student machine faces a dilemma of knowing when to follow the teacher and when to explore on its own. To this end, researchers from MIT and Technion, the Israel Institute of Technology, have developed an algorithm that automatically and independently determines when the student should mimic the teacher (known as imitation learning) and when it should instead learn through trial and error (known as reinforcement learning). https://shorturl.at/bOQX7

A model's performance decay can be due to different reasons, which is why it is important to update and retrain models consistently. Concept drift and model decay are not the same, and it is important to know the difference to understand the best way to keep the model’s accuracy - https://vzocca.substack.com/p/model-decay-vs-concept-drift-do-we

Want to learn more about Shapley values? What they are, how to use them, how to approximate them? Look no further - https://www.aidancooper.co.uk/approximating-shapley-values/

Five ways AI might destroy the world - https://www.theguardian.com/technology/2023/jul/07/five-ways-ai-might-destroy-the-world-everyone-on-earth-could-fall-over-dead-in-the-same-second

Party Planning Meets Graph Theory
To understand what a Ramsey number is, imagine you’re hosting a party.
How many people would you need to invite to guarantee that there will be a group of people who all know one another, or a group who are all strangers? You can encode this question in the language of graphs. Assign a vertex to each person. For n people, you get n vertices. Connect every pair of vertices with an edge. Color the edge red if the people in question know each other, and blue if they are strangers. https://rb.gy/qb2zr

“Quantum computing is a rapidly advancing field that has the potential to revolutionise the way we process and analyse data. In the realm of finance, quantum computing promises to provide powerful tools for financial modelling and risk assessment. However, it is not a straight path to success. This article will explore the benefits and harms of quantum computing and debate whether the financial sector will ever come to the point of requiring such a futuristic system.” https://shorturl.at/mpZ39

Defining good Key Performance Indicators (KPI) is a fundamental step for each business. Having good data is irrelevant if we do not know how to use it to track how close we are to achieving our goals - https://vzocca.substack.com/p/harnessing-the-power-of-data-driven

“Quantum computers owe their power in part to a phenomenon known as superposition, which effectively allows them to simultaneously explore many options that a classical computer would need to consider individually. But it’s not as simple as performing multiple calculations at once to save time. Checking the result of a superposition of choices never reveals a superposition of outcomes — rather, you only ever obtain one of the possible outcomes, each of which has a different probability. Quantum algorithms rely on the fact that contributions to these probabilities can interfere with each other like waves on the surface of a pond, boosting the probability of getting the right answer while reducing the probability of every other outcome.” https://rb.gy/tfrqo

Alchemy, the Philosopher's Stone and the Ten Capital Sins of businesses. Is Data Science the new Philosopher's Stone? Have we finally learnt how to transform base metals (disorganised data) into gold (useful insights)? But how do we prevent falling into the trap of misusing this process or to waste our resources? Read about the Ten Capital Sins of businesses when it comes to organising and using the data they are paying to acquire, store and analyse - https://vzocca.substack.com/p/alchemy-the-philosophers-stone-and

As companies collect ever-larger amounts of data, they need robust data governance to manage it responsibly. Data governance provides the policies and procedures to oversee data acquisition, usage, and deployment. With proper governance, companies can unlock the value in their data while respecting privacy and security. Rather than simply accumulating data, organizations must have a data strategy aligned with business objectives and guided by governance principles. Effective data governance enables trust and extract maximum benefit from data assets. - https://open.substack.com/pub/vzocca/p/data-governance-what-is-it-and-why

How far are we from true, human level, AGI? Why current LLMs are not yet the solution. Rather than creating a model of the world, they create a synopsis - https://aisupremacy.substack.com/p/how-far-are-we-from-agi

Towards AGI. What is missing?
To achieve AGI it seems likely we will need one or more of the following:
1. Online non-greedy planning
2. A world model
3. Reinforcement Learning Agent
- https://mark-riedl.medium.com/toward-agi-what-is-missing-c2f0d878471a

“Optimization problems like route planning, supplier management, and financial portfolio management are places where quantum’s unique ability to quickly find the optimal solution by analyzing huge amounts of heterogeneous data would work well,” she said. “Classical computers get overwhelmed by exponential calculations when it comes to these enormous amounts of data…AI and machine learning algorithms are perfect candidates for quantum processing.” https://shorturl.at/yDU19

Is it accurate to claim that having more data inevitably leads to better outcomes?Can data be deemed useful unconditionally?Should businesses strive to amass as much data as they can, assuming its intrinsic value?In my view, data holds value exclusively when it can be purposefully employed; otherwise, it could potentially result in escalated expenditures. Businesses must exercise discretion when determining the data to gather and retain, aligning their choices with their precise business requirements; otherwise, they might inadvertently transform into mere data hoarders with no clear business purpose - https://vzocca.substack.com/p/collect-all-data-possible