Access to the high-dimensional complexity resident in machine learning models comes at a cost. It is not being discussed much in public discussions but the ability to interrogate these complex information systems appears to bring an inevitable loss of control over the outputs. Generative systems in particular are prone to this inverse riddle: the more […]
Tag: Machine Learning
Bias in Machine Learning
Machine Learning amplifies existing signals in the data and also does so when we ourselves are quite unaware of that bias. Identifying biases and engaging them is, or has largely become, an art of post facto, retrodictive engagement. The futures technology generate are so often unacknowledged as recursively self-propagating compression waves, signals as an encoded […]
Induction
Context: Any Single Galaxy Reveals the Composition of an Entire Universe Such induction to general principles from instances is itself, also, a broader lesson regarding complexity. If recognisable (i.e. “real world”) properties of cosmological matter and energy distribution can be inferred from a single galaxy, in what other systems does such a depth of logical […]
AI is Huge, but…
AI is big. Huge. Consequential beyond anything we might have previously known was even on the technological radar of accelerating sociotechnical self-propagation but it does not and never could provide teleological closure. This is a reflection on foundational incompleteness and logical indeterminism. I think that it’s instructive to remark that the aspirational trajectories (as individuals, […]
Genetic Predestination
Context: Genetic paparazzi are right around the corner, and courts aren’t ready to confront the legal quagmire of DNA theft — I find that the unacknowledged (yet common enough) assumption that genes deterministically prescribe destiny as more or less isomorphic mappings of downstream biological development is unhelpfully shaping an already complex debate. Yes, genes shape […]
Marcus Aurelius once wrote that the very worst a person could do in life is to make themselves into a kind of blemish upon history and the world. We do not have to think very long or hard to find contemporary examples of just such an abhorrence and regrettably misanthropic wart upon all of history […]
Open AI have released the language model of GPT-3 for general use, with appropriate caveats. I have previously been skeptical of the ultimate utility of GPT-3 as a creative authorial or (useful) philosophically reflexive tool but discovered just now that if you feed it a sufficiently sophisticated prompt it actually throws back some interesting linguistic […]
Can a robot write a symphony?
Detective Del Spooner : Human beings have dreams. Even dogs have dreams, but not you, you are just a machine. An imitation of life. Can a robot write a symphony? Can a robot turn a… canvas into a beautiful masterpiece? Sonny : Can you? Movie: I, Robot, 2004. (After Asimov). A robot can write a symphony, it just can […]
Context: Guess who’s the biggest investor in self-driving tech? As a general (and perhaps self-evident) observation, note that the sum total of investment at $44 billion would achieve so much more if one company had those resources pooled but that, enigmatically, it is precisely the difference and competition which is driving the acceleration in this […]
Quantum Natural Language Processing
Context: Cambridge Quantum Releases World’s First Quantum Natural Language Processing Toolkit and Library We are moving away from chatbots that intermittently hurl absurdities and misinformation at (and as each other and) the world towards autonomous language-processing systems that reason and understand. Whether or not we get there is not as important as that we invoke […]
AI’s Poisoned Data
Context: DeepMind tells Google it has no idea how to make AI less toxic The raw data is indeed endemically and intermittently toxic. Filtering out venomous misanthropy from the dataset is a solution but this also requires a language model that does something substantively more than generate probabilistic sequences of linguistic artefacts sans comprehension. John […]
Distributed Intelligence
Intelligence is somewhat catastrophically problematised by our inability, from within the system, to ever fully capture or represent the system. From within language (and logic) we generate models as fantasies of complete and consistent truths that much more closely approximate to systems of belief that, similarly, simulate closure and completeness without ever being able to […]