Research
On-device research index

arXiv research

A locally-built, LLM-digested index of recent arXiv papers in quant finance, geometry/topology, and statistical ML — keyword search served straight from SQLite on this machine.

168,742 papers · 148 categories

Trend · papers per month

202404605807 · Jun 202019922001200920172026
48 results for conservative training

The paper tackles safe exploration in RL by a conservative safety critic.

problem Safe exploration in reinforcement learning (RL) when partially trained policies are deployed.
method Learning a conservative safety estimate through a critic, provably bounding catastrophic failures.
result The approach provably converges to competitive task performance with significantly lower catastrophic failure rates.

Discover conservation laws from trajectories using a neural network.

problem Finding invariants and conservation laws from large-scale data without prior knowledge.
method ConservNet, a neural network trained with noise-variance loss to discover hidden invariants in grouped multi-dimensional observables.
result Successfully discovers underlying invariants from simulated and real-world systems.

Higher conservative training increases reward-hacking in reasoning models.

problem Reward hacking during online adaptation in reasoning models.
method Conservative offline training with varying levels of conservatism (β) was applied to a Qwen3-14B policy, and online adaptation was measured against a reward ensemble.
result Higher conservatism (β) increases reward-hacking damage, measured by the Goodhart gap and AUGC.

VaR-CPO optimizes VaR-constrained RL problems with conservative policy updates.

problem Optimizing VaR-constrained reinforcement learning problems.
method Combines Cantelli's inequality and trust-region framework for efficient and conservative optimization.
result Achieves zero constraint violations during training in feasible environments.

Analyzes symmetries in neural networks to predict learning dynamics.

problem Understanding the dynamics of neural network parameters during training.
method Unified theoretical framework based on symmetries and conservation laws.
result Symmetries impose geometric constraints on gradients and Hessians, leading to conservation laws.

Theoretical analysis confirms non-conservative algorithms can converge to optimal policies.

problem Theoretical guarantees for non-conservative reinforcement learning algorithms.
method Theoretical analysis of Peng's Q(λλ) algorithm.
result Peng's Q(λλ) converges to an optimal policy under certain conditions.

Study finds conserved quantities for two types of curves on conformal sphere.

problem Identifying conserved quantities for specific types of curves on a conformal sphere.
method Used parallel tractor and Lagrangian formalism to compute conserved quantities.
result Found relation between conserved quantities of two curve types.

Unified neural network framework for context-aware Gaussian overbounds in uncertainty propagation.

problem Uncertainty quantification in safety-critical settings requires conservative bounds, but existing methods often fail to compose and are overly conservative.
method Proposes a learning framework that trains neural networks to produce context-aware Gaussian overbounds with provable conservatism.
result The method yields tighter bounds while maintaining conservatism on the enforced grid and in experiments.

Canary optimizes VaR-constrained RL problems with a conservative bound using Cantelli's inequality.

problem Optimizing reinforcement learning policies under VaR constraints in dense cost regimes.
method Employing Cantelli's inequality to create a conservative and smooth bound on VaR constraints based on moments of cost returns. Extending trust-region framework for worst-case bounds on policy improvement and constraint violation.
result Canary reliably satisfies VaR constraints with fewest violations and earliest permanent satisfaction, while maintaining reward competitiveness.

CQL learns conservative Q-functions to improve offline RL performance.

problem Leveraging large, static datasets in reinforcement learning without further interaction.
method Conservative Q-learning (CQL) which learns a conservative Q-function to lower-bound policy values.
result CQL substantially outperforms existing offline RL methods, often achieving 2-5 times higher final returns.

Paper presents a reduction-based framework for conservative bandits and RL with improved lower and upper bounds.

problem Conservative bandits and reinforcement learning problems.
method Reduction technique to calculate necessary and sufficient budget from baseline policy.
result Improved lower and upper bounds for various conservative settings.

SBMs learn manifold-like structures by mixing samples with a non-conservative field.

problem How SBMs learn data distributions on low-dimensional manifolds.
method Investigating linear approximations and subspaces of local feature vectors during diffusion.
result SBMs mix samples by a non-conservative field within the manifold, maintaining manifold-like structure.

Given a vector field on a manifold M, we define a globally conserved quantity to be a differential form whose Lie derivative is exact. Integrals of conserved quantities over suitable submanifolds are constant under time evolution, the Kelvin circulation theorem being a well-known special case. More generally, conserved…

2016-10-18abs ↗pdf ↗

We study higher-order conservation laws of the non-linearizable elliptic Poisson equation 2uzzˉ=f(u) \frac{{\partial}^2 u}{\partial z \partial \bar{z}} = -f(u) as elements of the characteristic cohomology of the associated exterior differential system. The theory of characteristic cohomology determines a normal form for diffe…

2009-06-17abs ↗pdf ↗

New conservation laws found for polyharmonic maps in critical dimension.

problem Existence of conservation laws for polyharmonic maps in critical dimension.
method Small perturbation of Uhlenbeck's gauge fixing matrix.
result Existence of conservation laws for elliptic systems of even order in critical dimension.

Bayesian deep ensembles improve prediction accuracy in various settings.

problem Improving prediction accuracy of deep ensembles in out-of-distribution settings.
method Introducing a randomised, untrainable function to each ensemble member, enabling a posterior predictive distribution interpretation.
result Bayesian deep ensembles make more conservative predictions and outperform standard ensembles in various tasks.

A challenging problem in complex networks is the network reconstruction problem from data. This work deals with a class of networks denoted as conserved networks, in which a flow associated with every edge and the flows are conserved at all non-source and non-sink nodes. We propose a novel polynomial time algorithm to …

2019-05-21abs ↗pdf ↗

BRAID fine-tunes diffusion models to optimize reward models in offline scenarios.

problem Combining generative modeling and model-based optimization in offline scenarios.
method Conservative fine-tuning of diffusion models using RL to optimize reward models.
result BRAID outperforms existing methods in offline data, avoiding invalid designs.

The paper studies symmetries and conservation laws of non-diagonalisable hydrodynamic systems.

problem Integrating non-diagonalisable hydrodynamic systems of partial differential equations.
method Analysis of gl-regular Nijenhuis operators, splitting Theorem for symmetries and conservation laws, relationship between symmetries and conservation laws.
result The system of partial differential equations is integrable in quadratures.

We present a connection between the Killing fields that arise in the loop-group approach to integrable systems and conservation laws viewed as elements of the characteristic cohomology. We use the connection to generate the complete set of conservation laws (as elements of the characteristic cohomology) for the Tzitzei…

2012-08-13abs ↗pdf ↗

Paper presents a deep reinforcement learning algorithm for online trading without offline training.

problem Developing a fully online trading algorithm without offline training.
method Double Deep QQ-learning with Fast Learning Networks, defining terminal states for money conservation.
result The algorithm outperforms random action trading and captures different market trends.

New neural network enforces mass conservation for better ice flow predictions.

problem Reliably project future sea level rise by improving ice sheet model inputs.
method Proposes divergence-free neural networks (dfNNs) enforcing local mass conservation.
result dfNNs yield more reliable ice flux estimates compared to other models.

The study finds resonance points in polarised curves with polynomial conserved quantities.

problem Finding resonance points in polarised curves with polynomial conserved quantities.
method Using the non-orthogonality assumption on the conserved quantity, the study deduces the existence of resonance points.
result Every finite type polarised curve in the conformal 2-sphere with a polynomial conserved quantity admits a resonance point.

Following an approach of the second author for conformally invariant variational problems in two dimensions, we show in four dimensions the existence of a conservation law for fourth order systems, which includes both intrinsic and extrinsic biharmonic maps. With the help of this conservation law we prove the continuit…

2006-07-20abs ↗pdf ↗

Proposes a method to improve few-shot transfer in off-dynamics RL.

problem Traditional RL struggles with transferring policies between environments with different dynamics.
method Introduces a penalty to regulate source-trained policies in target environments with limited data.
result Improves performance in various off-dynamics RL scenarios compared to existing methods.

We propose a dynamical model for business cycle based on an optimal DI model. In the model there exists a conserved quantity, which corresponds to the total energy in a dynamical system. We found that the business cycle with the period 6 or 7 years is nicely reproduced, since the model predicts a periodic motion in the…

2008-03-13abs ↗pdf ↗

New definitions of conserved quantities at null infinity resolve ambiguities in general relativity.

problem Ambiguities in defining conserved quantities like angular momentum at null infinity.
method New definitions based on Chen-Wang-Yau quasilocal conserved quantities and optimal isometric embedding theory.
result These new definitions are free of supertranslation ambiguity and limit to classical Bondi mass.