arXiv:2606. 08390v1 Announce Type: new Abstract: When a neural time-series model reports that one variable modulates another's effect on a target, is the discovered interaction a property of the data or an artifact of model flexibility?
arXiv:2608. 01587v1 Announce Type: cross Abstract: Machine-learning benchmarks often pair a label that aggregates a long temporal horizon with input observed through one or a few short windows.
arXiv:2606. 01092v1 Announce Type: cross Abstract: Supervised learning evaluates predictors through their input-output behavior.
arXiv:2606. 12138v1 Announce Type: cross Abstract: Sparse autoencoders (SAEs) are widely used to interpret neural network representations, but their utility depends on whether the learned features are reproducible across training runs.
The paper studies when joint-embedding predictive architectures (JEPAs) can recover underlying causal states from high‑dimensional observations. It introduces a latent variable model where observations arise from causal states with action‑conditioned dynamics, and proposes an information‑theoretic objective that maximizes conditional likelihood while preserving state entropy. The authors prove identifiability conditions—particularly sufficient action‑induced variation—and instantiate the objective as an action‑modulated Gaussian additive‑noise model (A‑JEPA), demonstrating theoretical and empirical success in synthetic and visual benchmarks.
arXiv:2604. 27967v2 Announce Type: replace Abstract: Background: We introduce StructGP, a continuous-time multi-task Gaussian process that couples process convolutions with differentiable structure learning to uncover a sparse, ordered directed acyclic graph (DAG) of inter-variable dependencies while preserving principled uncertainty.
arXiv:2510. 12744v2 Announce Type: replace-cross Abstract: We develop a unified statistical framework for softmax-gated Gaussian mixture of experts (SGMoE) that addresses three long-standing obstacles in parameter estimation and model selection: (i) non-identifiability of gating parameters up to common translations, (ii) intrinsic gate-expert interactions that induce coupled differential relations in the likelihood, and (iii) the tight numerator-denominator coupling in the softmax-induced conditional density.
arXiv:2502.04131v2 Announce Type: replace Abstract: The successful application of modern machine learning for time series classification is often hampered by limitations in quality and quantity of av...
The paper introduces ePID, an embedding-based approach that scales partial information decomposition (PID) to large symptom networks by compressing non‑focal symptoms into a low‑cardinality discrete embedding. Using a supervised Agglomerative Conditional Information Bottleneck (ACIB) embedding, ePID accurately recovers source‑unique, remainder‑unique, redundant, and synergistic components for each ordered source‑target pair across 83 real‑world datasets, outperforming 12 other embeddings. Applied to PHQ‑9 and the Interpersonal Reactivity Index, ePID reveals distinct patterns of redundancy and synergy that align with each instrument’s construction, demonstrating its ability to separate overlapping from interaction‑dependent information in symptom networks.
arXiv:2607. 26528v1 Announce Type: cross Abstract: Symbolic regression provides analytical expressions, but it is usually applied one output at a time.
arXiv:2609.23937v1 Announce Type: cross Abstract: Robust linear fits can resist response contamination yet remain too dense or unstable for useful global explanations. We propose penalized distillati...
arXiv:2607. 22313v1 Announce Type: cross Abstract: Estimating contemporaneous bidirectional interactions from observational data is difficult because each outcome is endogenous to the other, while flexible regressions may capture only reduced-form dependence.
arXiv:2603. 25414v4 Announce Type: replace-cross Abstract: A prevailing assumption in machine learning is that model correctness must be enforced after the fact.