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.
Large language models can't efficiently reason conditionally in a distribution-free setting.
Achieving artificial visual reasoning - the ability to answer image-related questions which require a multi-step, high-level process - is an important step towards artificial general intelligence. This multi-modal task requires learning a question-dependent, structured reasoning process over images from language. Stand…
We introduce a general-purpose conditioning method for neural networks called FiLM: Feature-wise Linear Modulation. FiLM layers influence neural network computation via a simple, feature-wise affine transformation based on conditioning information. We show that FiLM layers are highly effective for visual reasoning - an…
New method for LLMs to learn reasoning by optimizing latent variables.
Transformers with CoT don't enhance reasoning power across all tasks.
Alpha-R1 uses LLMs to reason about economic factors and news for better alpha screening.
A new method for math reasoning that allows for iterative correction.
VTA combines verbal and latent reasoning for accurate stock time-series forecasts.
This work identifies and mitigates reasoning shortcuts in Neuro-Symbolic models.
We prove that, under reasonable conditions, odd co-dimension Riemannian foliations cannot occur in positively curved manifolds.
The problem of replicating the flexibility of human common-sense reasoning has captured the imagination of computer scientists since the early days of Alan Turing's foundational work on computation and the philosophy of artificial intelligence. In the intervening years, the idea of cognition as computation has emerged …
A conformal procedure improves CoT reasoning by aggregating reasoning paths and calibrating abstention rules.
Hybrid deep architectures with reasoning layers show promising convergence and generalization properties.
Dynamic abstention improves LLM accuracy by selectively terminating unpromising reasoning.
New causal versions of MaxEnt and PIR avoid paradoxical probability updates.
REALFIN benchmarks financial reasoning by removing implicit assumptions, revealing model weaknesses.
New approach to counterfactual reasoning avoids demographic interventions.
Adaptive Nucleus Truncation Improves Long-Form Reasoning
Self-supervised method improves representation learning for better accuracy.
Unified CI test for categorical and ordinal data maintains power in high dimensions.
For autonomous vehicles (AVs) to behave appropriately on roads populated by human-driven vehicles, they must be able to reason about the uncertain intentions and decisions of other drivers from rich perceptual information. Towards these capabilities, we present a probabilistic forecasting model of future interactions b…
We present a stochastic numerical method for solving fully non-linear free boundary problems of parabolic type and provide a rate of convergence under reasonable conditions on the non-linearity.
In the framework of standard static space times, we state a family of sufficient or necessary conditions for a set of physically reasonable energy and convergence conditions in relativity and related theories. We concentrate our study on questions about the sub-harmonicity of the warping function, the scalar curvature …
AGENTICAITA uses AI agents to autonomously trade markets without human intervention.
In this paper, we prove a Morse index theorem for the index form of even order linear Hamiltonian systems on the closed interval with reasonable self-adjoint boundary conditions. The highest order term is assumed to be nondegenerate.
Bayesian framework integrates spectral deconvolution with expert reasoning for robust peak estimation.
Causal reasoning has been an indispensable capability for humans and other intelligent animals to interact with the physical world. In this work, we propose to endow an artificial agent with the capability of causal reasoning for completing goal-directed tasks. We develop learning-based approaches to inducing causal kn…
Humans are capable of attributing latent mental contents such as beliefs or intentions to others. The social skill is critical in daily life for reasoning about the potential consequences of others' behaviors so as to plan ahead. It is known that humans use such reasoning ability recursively by considering what others …
New approach to counterfactual reasoning in AI and psychology.
We present a multisymplectic formulation of the Yang--Mills equations. The connections are represented by normalized equivariant 1-forms on the total space of a principal bundle, with values in a Lie algebra. Within the multisymplectic framework we realize that, under reasonable hypotheses, it is not necessary to assum…
The paper analyzes finite-time singularities in Spin(7)-structure flows using Shi-type estimates.
The paper develops a theory for iterative self-improvement of models, proving conditions for better performance with easy-to-hard curricula.
AdaBoost is one of the most popular ML algorithms. It is simple to implement and often found very effective by practitioners, while still being mathematically elegant and theoretically sound. AdaBoost's interesting behavior in practice still puzzles the ML community. We address the algorithm's stability and establish m…
Causal knowledge is vital for effective reasoning in science, as causal relations, unlike correlations, allow one to reason about the outcomes of interventions. Algorithms that can discover causal relations from observational data are based on the assumption that all variables have been jointly measured in a single dat…
Power-SMC reduces inference latency for training-free LLM reasoning.
We showed how to use trained neural networks to perform Bayesian reasoning in order to solve tasks outside their initial scope. Deep generative models provide prior knowledge, and classification/regression networks impose constraints. The tasks at hand were formulated as Bayesian inference problems, which we approximat…
New research challenges the independence assumption in neurosymbolic learning, leading to overconfident predictions and unrepresentable uncertainty.
In an independence model, the triplets that represent conditional independences between singletons are called elementary. It is known that the elementary triplets represent the independence model unambiguously under some conditions. In this paper, we show how this representation helps performing some operations with in…
LaTRO optimizes latent reasoning in LLMs without external reward.
Neural networks are among the most accurate supervised learning methods in use today, but their opacity makes them difficult to trust in critical applications, especially when conditions in training differ from those in test. Recent work on explanations for black-box models has produced tools (e.g. LIME) to show the im…
Auto-CEI improves LLM reasoning by balancing assertiveness and conservativeness.
A framework isolates VQA reasoning from perception for better model evaluation.
UCPO improves diversity in reinforcement learning models, maintaining high accuracy.
We show that under reasonable conditions, the spines of the handlebodies of a strongly irreducible Heegaard splitting will intersect a closed ball in a graph which is isotopic into the boundary of the ball. This is in some sense a generalization of the results by Scharlemann on how a strongly irreducible Heegaard split…
It has been shown elsewhere that a strongly irreducible Heegaard splitting surface Q of a compact orientable 3-manifold M can, under reasonable side conditions, intersect a ball or a solid torus in M in only a few possible ways. Here we extend those results to describe how Q can intersect a handlebody in M.
This work improves motion planning for quadcopters by learning and reasoning about controller performance.
Inferring new facts from existing knowledge graphs (KG) with explainable reasoning processes is a significant problem and has received much attention recently. However, few studies have focused on relation types unseen in the original KG, given only one or a few instances for training. To bridge this gap, we propose Co…
Transformers learn multi-step reasoning through gradient descent.