Experimental Discovery

Extracted Attributes

• Goal

  Innovation, growth, and long-term profitability

• Key Proponent

  Charles Koch

• Core Framework

  Principal-Based Management

• Primary Methodology

  Bottom-up empowerment and hypothesis testing

• Organizational Model

  Republic of Science

• Scientific Application

  Pivotal in establishing the Standard Model of particle physics (e.g., W and Z bosons)

• Internal Testing Facility

  Koch Labs

• Associated Economic Concept

  Creative Destruction

Timeline

• Charles Koch analyzes principles of Comparative Advantage and Division of Labor during his early career at Arthur D. Little and MIT. (Source: Document 55c5ba4c-d2b8-4c11-b9ca-1afc406bc189)

  1960-01-01

• Experimental discovery of W and Z bosons, pivotal for the Standard Model of particle physics. (Source: Wikipedia)

  1983-01-25

• Koch Industries acquires Georgia-Pacific, subsequently replacing its 51-story hierarchy with meritocratic systems based on experimental discovery. (Source: Document 55c5ba4c-d2b8-4c11-b9ca-1afc406bc189)

  2005-12-23

• Chase Koch establishes Koch Disruptive Technologies to champion permissionless innovation and experimental discovery in tech. (Source: Document 55c5ba4c-d2b8-4c11-b9ca-1afc406bc189)

  2017-11-01

• Publication of 'Experimental Design for Scientific Discovery' dissertation by Quan Minh Nguyen at Washington University in St. Louis. (Source: Web Search Result)

  2024-07-02

• Last update to the Product Talk Discovery Glossary regarding experiments and assumption testing. (Source: Web Search Result)

  2025-10-25

W and Z bosons

In particle physics, the W and Z bosons are vector bosons that are together known as the weak bosons or more generally as the intermediate vector bosons. These elementary particles mediate the weak interaction; the respective symbols are W+, W−, and Z0. The W± bosons have either a positive or negative electric charge of 1 elementary charge and are each other's antiparticles. The Z0 boson is electrically neutral and is its own antiparticle. The three particles each have a spin of 1. The W± bosons have a magnetic moment, but the Z0 has none. All three of these particles are very short-lived, with a half-life of about 3×10−25 s. Their experimental discovery was pivotal in establishing what is now called the Standard Model of particle physics. The W bosons are named after the weak force. The physicist Steven Weinberg named the additional particle the "Z particle", and later gave the explanation that it was the last additional particle needed by the model. The W bosons had already been named, and the Z bosons were named for having zero electric charge. The two W bosons are verified mediators of neutrino absorption and emission. During these processes, the W± boson charge induces electron or positron emission or absorption, thus causing nuclear transmutation. The Z boson mediates the transfer of momentum, spin and energy when neutrinos scatter elastically from matter (a process which conserves charge). Such behavior is almost as common as inelastic neutrino interactions and may be observed in bubble chambers upon irradiation with neutrino beams. The Z boson is not involved in the absorption or emission of electrons or positrons. Whenever an electron is observed as a new free particle, suddenly moving with kinetic energy, it is inferred to be a result of a neutrino interacting with the electron (with the momentum transfer via the Z boson) since this behavior happens more often when the neutrino beam is present. In this process, the neutrino scatters off the electron (via exchange of a boson), transferring some of the neutrino's momentum to the electron.

Web Search Results

• Experimental Discovery

  To encourage Experimental Discovery, we don’t penalize well-planned experiments that fail since they fuel the necessary flow of small and frequent bets that generate discovery and learning. This is vital to innovation, growth, and long-term profitability. It is also motivating, as experimenting to discover new ways to create value makes work more interesting and exciting. [...] A well-designed experiment starts with a hypothesis and the goal of learning whether it is valid. If done properly, it leads to new knowledge that brings about change, even if our assumptions or hypotheses are disproven. We learn even more when we explore a range of possibilities that includes the areas of greatest uncertainty and potential. Confusing as it might seem, failure and getting results are not mutually exclusive. As Einstein is believed to have observed, “Failure is success in progress.” A failed but well-designed experiment is valuable if it generates lessons that lead to positive results. A true failure is a failure to learn because of poorly planned or impulsive action. [...] Skip Navigation Toggle Mobile Menu # Principles in Brief ### Experimental Discovery As commentator George Will reminded us: “The future has a way of arriving unannounced.” In our rapidly changing world, competitors are constantly improving, and what customers value is constantly changing. No matter how superior a company’s knowledge, products, and services, it cannot stay in business unless it makes improvements and innovations at least as fast as its most effective competitors. Doing this successfully requires that a business apply Experimental Discovery and Creative Destruction to its vision, strategies, products, services, and methods. All businesses must continually innovate, which usually involves numerous changes in direction, leading to the discovery of new paths.

• "Experimental Design for Scientific Discovery" by Quan Minh Nguyen

  Experimental design offers an elegant model of many problems where one navigates within a vast search space seeking data points with certain characteristics. A multitude of applications in science and engineering fall under this umbrella, with drug and materials discovery being prime examples. The experimental design approach maintains a probabilistic model of the search space, and uses Bayesian decision theory accounting for this model to guide the accumulation of observed data to maximize an experimentation objective of interest. This dissertation explores Bayesian optimization and active search, two realizations of the experimental design framework that model discovery tasks. While existing solutions are available for these two problems under conventional settings, there are important [...] problems under conventional settings, there are important scenarios to which these solutions cannot be readily applied, namely those of high dimensions or with multiple data sources, objectives that favor diversity in the collected data, and settings where efficient policy computation is crucial such as real-time systems and large-scale databases. We address these gaps, putting forward optimization and search policies with competitive empirical performance under their respective settings. The algorithmic solutions in our works provide practitioners with the tools to tackle a broad range of experimental design tasks, and ultimately advance machine learning-aided scientific discovery efforts. [...] ## Committee Chair Roman Garnett ## Degree Doctor of Philosophy (PhD) ## Author's Department Computer Science & Engineering ## Author's School McKelvey School of Engineering ## Document Type Dissertation ## Date of Award 7-2-2024 ## Language English (en) ## DOI ## Recommended Citation Nguyen, Quan Minh, "Experimental Design for Scientific Discovery" (2024). McKelvey School of Engineering Theses & Dissertations. 1062. The definitive version is available at Download opens in new window") ### Included in Computer Sciences Commons ## Share COinS #### DOI ## Search Advanced Search Notify me via email or RSS ## Links Graduate Student Services, McKelvey School of Engineering ## Browse Collections Disciplines Authors ## Author Corner

• Experiments | Definition and Overview | Product Talk

  Teams rarely have time to run real experiments in discovery. They can run large-scale randomized controlled experiments, also known as A/B tests, after they build something, but that's only if they have enough traffic. And this is not the best discovery activity. Teams don't want to do all the work to build something before they learn they built the wrong thing. This is why assumption testing makes more sense during discovery—assumption tests are faster and take less work. Assumption tests are used in discovery when trying to decide between ideas. Experiments are used to measure the impact of what teams built. ## When should teams use experiments? [...] ## When should teams use experiments? While experimentation is becoming more pervasive in product development, teams need to be smart about when and how they use experiments. Teams should use lighter-weight assumption tests during discovery and save full experiments for measuring the impact of what they deliver. Learn more: - Assumption Testing: Everything You Need to Know to Get Started - Hypothesis Testing Related terms: - Assumption Testing - Hypothesis - Prototyping - Discovery ← Back to Discovery Glossary Last Updated: October 25, 2025 ## Make better product decisions. Get actionable insights delivered straight to your inbox. [...] Product Talk Sign in Subscribe ## Make better product decisions. Get actionable insights delivered straight to your inbox. # Experiments ## What are experiments? Experiments are structured activities designed to support or refute a hypothesis by testing whether a specific change has the expected impact. In product discovery, the term "experiments" typically refers to large-scale randomized controlled experiments (like A/B tests) that measure the impact of something after it's been built. ## What's the difference between experiments and assumption tests?

• Adaptive experiments: Machine learning helps scientific discovery

  Machine learning can help scientists design experiments. Scientific discovery relies on experiments that build our understanding of natural phenomena, and traditionally has been based on trial and error. Depending on the goal, different machine learning strategies can be used for adaptive experiments: active learning, maximising information gain, Bayesian optimisation, bandit approaches, and reinforcement learning. Cheng Soon Ong, machine learning scientist at CSIRO, Australia, shows how these techniques can be applied to biology in particular. [...] In his 2013 study, machine learning was used to make sense of complex biological systems. By using information gain approaches, experimental decisions can be made a priori based on accumulated knowledge. Ong performed time point selection for experiments on glucose tolerance. He also identified the most informative data to elucidate the Target-of-Rapamycin (TOR) signalling pathway. Efficient design, therefore, improves the loop between scientific modelling and experimentation as it sheds light on transcription factor dynamics linked to metabolism. This innovation can be applied to other models in systems biology and will eventually accelerate scientific discovery. [...] ### Why do we need machine learning for scientific discovery? Scientific discovery is characterised by both data and knowledge: going from data to knowledge is defined as an observation, while going from knowledge to data is called experimentation. Adaptive experiments optimise this loop between data and knowledge. In an experiment, data collection may be expensive, meaning that we need to prioritise the measurement of informative data. In addition, experiments are often characterised by time-consuming iterative cycles, which could be partially automated while retaining the same performance. Thus, it is important to design the best possible sequence of experiments to observe the best possible data and to quickly discover new knowledge.

• Autonomous Scientific Discovery – Computing Sciences

  Generating scientific hypotheses and data has traditionally relied on laborious human effort. However, the ongoing revolution in artificial intelligence (AI) and robotics, combined with advanced networking and computing hardware, offers the promise of automated scientific discovery. Advanced algorithms, simulations coupled with experiments, and next-generation networking and computing infrastructure are essential for automating scientific discovery across the DOE mission space. At Berkeley Lab, our work includes developing AI-based algorithms for real-time steering of beamlines at DOE light sources, specifying targets for telescopes, and designing molecules with desired properties. We are also building data-management workflows that enable seamless integration and access throughout the [...] Connecting experimental infrastructure to high-performance computing is becoming increasingly critical. This integration provides a foundation for developing self-driving and self-guiding facilities and supports autonomous scientific discovery. At Berkeley Lab, our efforts are focused on defining automated architectural models that facilitate streaming data from experimental facilities, enhance resilience, and integrate tools for sharing, searching, and analyzing data. These advancements are designed to foster more productive and reproducible science. Together, these innovations will enable the automation of diverse DOE science experiments and accelerate the pace and efficiency of scientific discovery. ### Our Research Pillars: