Scars - A Sidebar Journey into LLM Philosphical Models

 

"Data is key to developing AI."(1)

OrbMB's mission is to help optimize compute and dataset production, so we watch progress in adjacent spaces. So learned that "Alignment" thought leader Jan Leike (2) resigned from OpenAI. 

Alignment is the structure of belief that: "We need scientific and technical breakthroughs to steer and control AI systems much smarter than us."(3)

Reviewing Leike's work led to the work of Mark Hutter (4) (who seeks to align maths, philosophy and particle physics); and this got me pondering the nature of social values that are at the heart of the philosophy of a training model build. 

Are LLM builders first designing the philosophy, to define the nature of the build, and only then starting the model build? If not, consider using a familiar analogy - the nature of schooling - the mechanism of the teaching that each of us experienced as children, and what results from the mechanism?

A) Heartless: Corporal Punishment = What results?
B) Thoughtful (Disciplined without punishment): Hybrid Training = ?
C) Heartful: Imaginative Play = ?

Last week at the gym, a buddy and me were joking. He is shredded; and I am not, and have scars; and joked that "One day, I will have six-pack scars." His response? "We are what our scars make us."

Everyone of us has lived all three methods mixed together.

Everyone of us carries those lessons along the course of our lives.

Could it be that a "philosophy of heartlessness" is the risk? Could it be that a "philosophy of heartfulness" is what will align self-aware human and self-aware AI values?

-----------------------------

(1) William “Bill” Streilein, 2023 Data, Analytics, and AI Adoption Strategy | 11 April 2024
https://www.youtube.com/watch?v=d4NcRQRwqIo  

(2) Jan Keike: https://jan.leike.name/

(3) https://openai.com/index/introducing-superalignment/ 

(4) Mark Hutter: http://hutter1.net/

 

 

 

Navigating the Value Conundrum - Azcoitia (2023)

The next several blog posts will discuss the current state of the art of data valuation.

Azcoitia, S.A., Towards a Human-Centric Data Economy, Ph.D dissertation; Telematics Engineering, Universidad Carlos III de Madrid https://doi.org/10.48550/arXiv.2111.04427 and https://sandresazcoitia.com/2023/04/24/towards-a-human-centric-data-economy/

 

 

 

 

 

 

Unravelling the Complex Web of Data Valuation: 

In the ever-evolving landscape of data-driven economies, understanding the true value of data has become a paramount challenge. This 2023 study is a comprehensive exploration that delves into the intricacies that arise when determining the worth of data assets; whilst navigating the dynamic realm of commercial data marketplaces. The study sheds light on the data value chain, and the nuances of trading data assets through the internet, through a meticulous examination of market dynamics.

• Part (1) explores the data value chain and the trading of data assets.
• Part (2) reports development and execution of a measurement study.
• Part (3) reports development of novel algorithms and methods to streamline data transactions.
  
• Part (4) concludes, presents new research topics, and proposes policy changes.

Part 1: (p.3) Begins by dissecting the data value chain and delving into the trading mechanisms of data assets facilitated by the internet. A detailed survey and analysis of commercial data marketplaces and vendor strategies makes up this section. The subsequent market review, starting on page 29, is highly recommended for a comprehensive understanding of the current landscape.

Part 2: Reports development and execution of a measurement study that estimates the value of data and the setting the price of a dataset; and uses this to analyse the behaviour of market data prices. This activity is aimed at estimating value to establish a pricing framework. 

Part 3: Azcoitia reports development of a framework of “algorithms and tools to reduce the complexity”, improve market efficiency, and improve buyer profitability. The framework is used to analyze the behaviour of market data prices; providing valuable insights into the dynamics of this complex ecosystem. 

Part 4: Discusses the findings and includes proposals for policy changes arising from the author's observation that:

• The elusive nature of data as an economic good has proven to be a central problem (pages 12-16). 
• Estimating value has proven to be challenging, leading to seemingly contradictory estimations. 
  
• The quality of data emerges as a crucial production factor, comparable to traditional factors such as land, capital, labour, or infrastructure.

The author emphasizes the critical need for a nuanced approach to assess the true worth of data in economic terms. Intricacies of pricing strategies are unravelled in pages 17-18, providing a comprehensive review of various approaches such as usage-based, subscription-based, and package-pricing.

In conclusion, the study examines the complexities that are part of estimating value and proposes algorithms and tools to reduce this complexity, enhance market efficiency, and improve buyer profitability. The author advocates for policy changes to foster a more transparent and adaptive data marketplace. 

As industries grapple with the evolving data landscape, this study serves as a crucial guide for navigating the intricacies of today's data markets.

- Summary notes re-written with the assistance of a paid ChatGPT account.