Writing
Identification of highly boosted decays with the ATLAS detector using deep neural networks
This thesis introduces two jet tagging algorithms to identify highly boosted decays using the ATLAS detector at the LHC. Based on the Deep Neural Network (DNN) architecture, the first algorithm's performance is comparable to an existing algorithm designed for highly boosted decays. The DNN jet tagger is also multifunctional and highly effective for identifying decays. Notably, it displayed enhanced rejection rates for background -jets. The second algorithm leverages an Adversarial Neural Network (ANN) architecture for mass-decorrelated classification. While it exhibited a slight performance decrease compared to the DNN-based tagger, it demonstrated a reduction in mutual information between the mass feature and scalar discriminant metric, substantiating its capability for mass-decorrelated jet identification.
Why Code Generating Models Got Good
A literature review tracing how code-generating language models improved as their architectures captured more context. The arc runs from Hindle et al.'s naturalness hypothesis, through n-grams, RNNs and LSTMs, to transformer models like Codex.
Unlimited Associative Learning and the Natural Kind Status of Phenomenal Consciousness
Ginsburg and Jablonka's unlimited associative learning is best read not as phenomenal consciousness itself but as a nomological cluster of evidential properties, in the sense of Shea and Bayne's natural kind methodology. The two frameworks fit together, and the synthesis licenses a broader empirical research programme into non-human consciousness.
Transcendental Idealism and QBism: Knowing through Appearances in Kant and Quantum Theory
We read Kant's transcendental idealism as primarily an epistemic project and side with Allison's two-aspect interpretation. On that reading, QBism is the interpretation of quantum physics Kant would most plausibly endorse, since it treats the wavefunction as subjective belief while keeping observable outcomes empirically real.
No rules of induction: two responses to Hume's problem
Hume's problem asks what justifies induction when no universal schema can do the work without circularity. Samir Okasha and John Norton both respond by denying that inductive inference needs such a schema. This essay compares their approaches and argues they converge on a single no-rules account, in which licensing facts derived from prior empirical inquiry justify particular inferences.
Chronogeometric Fatalism and the Philosophy of Time in Special Relativity
Special relativity collapses Newton's separate space and time into a single Minkowski geometry. The most natural metaphysical reading is chronogeometric fatalism, on which all points of space-time, past, present and future, are equally real.
Cold Gas Kinematics in the z=4 Submillimeter Galaxy AzTEC-1
AzTEC-1 is an extreme starburst galaxy at z=4.3, seen as it was when the universe was 1.5 billion years old. This thesis compares two cold gas tracers, [CII] and CO, with kinematic modelling of ALMA data to probe rotation, dispersion, and a curious non-corotating clump.
Detecting Dark Matter: The Multi-Messenger Approach
Dark matter's gravitational fingerprints have been visible since Zwicky's 1933 cluster measurements, yet its composition has eluded direct identification for nearly a century. This essay surveys the three complementary detection strategies (direct scattering experiments, collider production, and indirect searches for annihilation products) that together comprise the modern multi-messenger hunt, and argues that indirect detection of WIMPs offers the most economical and versatile route to discovery.
Fitting the GALPROP Model to AMS-02 Lithium Data via the Diffusion Term
Fitting Stanford's GALPROP cosmic ray propagation model to AMS-02 lithium data through the diffusion term alone. The result is a poor fit above the 200 GV break, consistent with the lithium anomaly.
Energy Transport in the Sun: From Models to Observables
The Sun is the only star whose interior we can probe in detail. This essay traces how radiative diffusion, proton-proton fusion, and the mixing-length approximation for convection each tie back to an observable characteristic we can measure.