Online gradient descent can simulate complex computations.
problem Understanding the fine-grained behavior of online gradient descent is hard.
method Proving online gradient descent can encode arbitrary polynomial-space computations.
result It is impossible to reason efficiently about the fine-grained behavior of online gradient descent under weak complexity-theoretic assumptions.
Study of embeddings avoiding certain tangent patterns using polynomial spaces.
problem Classifying embeddings avoiding specific tangent patterns.
method Introduce equivalence relation (quasitopy) and use spaces of polynomials as Grassmannians.
result Quasitopy classes of Θ-constrained embeddings stabilize as degree increases.
This paper continues the work of our previous paper [8], where we generalize kth-powers of the Euclidean Dirac operator D_x to higher spin spaces in the case the target space is a degree one homogeneous polynomial space. In this paper, we reconsider the generalizations of D_x^3 and D_x^4 to higher spin spaces in the ca…
Classifies homomorphisms between specific braid groups.
problem Classifying homomorphisms between braid groups.
method Complete classification through recursive approach.
result Recursive classification of homomorphisms between braid groups.
PDEs constrain smooth functions in neural networks.
problem Understanding functions computable by neural networks.
method Analyzing smooth hierarchical functions via PDEs.
result Established algebraic PDEs for smooth functions.
Lower bounds show density estimation requires linear samples or query time.
problem Statistical-computational trade-offs in density estimation.
method Lower bound analysis on data structures.
result Lower bounds demonstrate statistical-computational trade-offs for density estimation.
New methods distill dynamical system equations from data.
problem Discovering governing equations of dynamical systems from data.
method Sparse regression, LASSO, dual LASSO optimization, STRidge algorithm.
result Improved accuracy and stability in learning dynamical system equations.
This paper studies a particular class of higher order conformally invariant dif- ferential operators and related integral operators acting on functions taking values in particular finite dimensional irreducible representations of the Spin group. The differential operators can be seen as a generalization to higher spin …
POUnets combine partitions of unity and monomials for efficient deep learning.
problem Efficiently approximating functions with deep neural networks in high dimensions.
method Integrates partitions of unity and monomials into neural network architecture.
result POUnets achieve hp-convergence for smooth functions and outperform MLPs for discontinuous functions.
Polynomial delay algorithm tests causal models with hidden variables.
problem Testing causal models with hidden variables in polynomial delay.
method c-component local Markov property (C-LMP) and polynomial delay algorithm.
result First algorithm for poly-delay testing of CIs in causal graphs with hidden variables.
The paper studies geometric structures of polynomial spaces.
problem Understanding the geometric and combinatorial structures of polynomial spaces.
method Introducing and analyzing finite piecewise Euclidean cell complexes.
result The branched rectangle and annulus complexes are homeomorphic to specific polynomial spaces.
Quantum computing speeds up multi-period asset allocation.
problem High computational complexity in classic computing for multi-period asset allocation.
method Applied quantum computing to simulate multi-asset portfolio using historic data.
result Quantum computing offers significant advantages over classical computing in finance.
Hybrid approach reduces computation time and decoding complexity.
problem Straggling servers in distributed computing.
method Coded partial gradient computation (CPGC) that balances gradient accuracy and completion time.
result Reduces both computation time and decoding complexity.
The paper analyzes the pricing of a new compute futures asset.
problem Uncertainty in AI adoption and pricing of compute capital.
method An asset-pricing framework for compute futures, including synthetic futures pricing.
result Preliminary evidence suggests a positive compute risk premium.
Quantum computing offers energy savings over classical computing.
problem Energy efficiency in computing services.
method Cournot competition model constrained by energy usage.
result Quantum computing firms can outperform classical counterparts in energy efficiency.
The paper introduces reservoir computing models for complex systems.
problem Modeling complex engineering systems using nonlinear autoregression.
method Introduces reservoir computing with output feedback as stationary and ergodic infinite-order nonlinear autoregressive models.
result Demonstrates versatility of classical and quantum reservoir computers in modeling synthetic and real data.
Knot theory applied to quantum computing models.
problem Using knot theory for quantum computing models.
method Exploring knot theory applications in quantum computing.
result Knot theory introduces topological concepts to quantum computing.
This work makes neural sequence models more efficient by controlling computation.
problem Fixed compute for all examples in neural networks.
method Conditional computation to adapt compute to example complexity.
result Conditional Computation Transformer (CCT) improves efficiency and performance.
Defines computable learning for binary classification over metric spaces.
problem Defines computable PAC learning for binary classification over computable metric spaces.
method Provides sufficient conditions for ERM learners to be computable and bounds the strong Weihrauch degree of an ERM learner.
result Gives a hypothesis class that does not admit any proper computable PAC learner with computable sample function.
Automatic computation speeds up crosscap number calculation for alternating knots.
problem Computing crosscap numbers for alternating knots efficiently.
method Introduced an automatic computation with complexity O(E3). result Crosscap numbers of alternating knots can be computed in O(E3) time. TKFT models computation via smooth vector fields, simulating functions in a single dynamical step.
problem Modeling computation in a single step.
method Established Topological Kleene Field Theory (TKFT) as a new model of computation.
result Any computable function can be simulated in a single go of a dynamical system.
Predicts and classifies computational jobs for efficient resource allocation in cloud centers.
problem Efficiently scheduling and assigning resources to computational jobs in cloud centers.
method Applied LSTM neural network for job arrival prediction and BIRCH clustering for job classification.
result Improved accuracy in predicting and classifying computational jobs compared to existing methods.
Stochastic reservoir computing is shown to be a universal approximator.
problem Theoretical justification for using stochastic reservoirs in machine learning.
method Investigated stochastic reservoir computing using probabilities of reservoir states as readout.
result Stochastic reservoir computers are universal approximating classes.
Machine learning impacts computational math, offering new functions approximations.
problem Machine learning's black box nature hinders further progress in computational math.
method Analyzes machine learning's impact on computational math and vice versa.
result Integrating computational math with machine learning can enhance both fields.
Method for computing Khovanov homology of tangles.
problem Limited explicit computational studies of Khovanov homology for tangles.
method Arc reduction approach to compute Khovanov homology.
result Derived and computed Poincaré polynomials for simple and complex tangles.
This paper simplifies computing higher-order U-statistics efficiently.
problem The inefficiency of computing higher-order U-statistics in practice. method Decomposition, connection to Einstein summation, and treewidth-based complexity estimate.
result A new, more efficient algorithm to compute U-statistics. Quantum reservoir computing tackles noisy quantum computers for temporal tasks.
problem Efficiently process input sequences on noisy quantum computers.
method Quantum reservoir computing using dissipative quantum dynamics.
result Small and noisy quantum reservoirs can handle high-order nonlinear temporal tasks.
Survey on computational models in dynamical systems, including new universality concepts.
problem Understanding the relationship between computational models and dynamical systems.
method Review of recent works on Turing universality, Topological Kleene Field Theories, and dynamical bordisms.
result Introduction of new perspectives on computability through dynamical systems.
Computations for prime knots up to 11 crossings.
problem Computing HOMFLY homology for prime knots.
method Direct computations for all prime knots up to 11 crossings.
result HOMFLY homology determined for all prime knots up to 11 crossings.
Blockchain as a Service offers a secure, decentralized computing solution.
problem Lack of transparency, security, and privacy in cloud computing.
method Decentralized cooperative computing process using blockchain, homomorphic encryption, and SDN.
result Performance evaluated via different scenarios in simulations.
Survey on quantum computing and neural networks.
problem Understanding and comparing quantum computing and neural networks.
method Introduction to quantum computing concepts, explanation of quantum computing paradigms, and analysis of quantum neural networks.
result Current state-of-the-art in quantum neural networks.
This paper optimizes mobile device computation offloading using deep reinforcement learning.
problem Optimizing computation offloading for mobile devices in virtual edge computing systems.
method Modeling the problem as a Markov decision process and using double deep Q-networks for learning optimal offloading policies.
result The proposed algorithms significantly improve computation offloading performance.
New methods improve Reservoir Computing for chaotic time series prediction.
problem Chaotic time series prediction in Reservoir Computing.
method Established Recurrent Kernel limit, introduced Structured Reservoir Computing.
result Structured Reservoir Computing is faster and more memory-efficient.
Quantum computing promises faster bioinformatics, but challenges remain.
problem Efficient bioinformatics processing and drug discovery.
method Quantum algorithms for optimization, simulation, and machine learning.
result Quantum computing can significantly speed up bioinformatics tasks.
Computer-generated proofs led to a mathematical result.
problem Discovering a mathematical result through computer-generated proofs.
method Combining computer-generated, human-readable proofs with mathematical abstraction.
result Abstracted lemma leading to an interesting mathematical result.
This work shows how to compute subderivatives efficiently without errors.
problem Inefficient and incorrect computation of subderivatives in ML libraries.
method Developed a method to compute provably correct generalized subderivatives at a cost close to the function itself.
result Provable correct generalized subderivatives can be computed at a cost within a factor of 6 of the function itself.
A DRL approach optimizes computation offloading in MEC systems for mobile users.
problem Optimizing computation offloading in MEC systems with mobile users and stochastic task arrivals.
method Deep Deterministic Policy Gradient (DDPG) for decentralized dynamic computation offloading.
result The DDPG-based strategy outperforms conventional strategies in terms of computation cost and power-delay tradeoff.
We present two paradigms relating algebraic, topological and quantum computational statistics for the topological model for quantum computation. In particular we suggest correspondences between the computational power of topological quantum computers, computational complexity of link invariants and images of braid grou…
As inductive inference and machine learning methods in computer science see continued success, researchers are aiming to describe ever more complex probabilistic models and inference algorithms. It is natural to ask whether there is a universal computational procedure for probabilistic inference. We investigate the com…
Parallelizes feedforward computation using nonlinear equation solving.
problem Sequential nature of feedforward computation limits parallelization.
method Frame feedforward computation as solving nonlinear equations; use Jacobi or Gauss-Seidel methods for parallel updates.
result Accelerates feedforward computation with reduced parallelizable iterations.
Study error bounds in evaluating distributional computational graphs.
problem Error analysis in evaluating graphs with inputs as probability distributions.
method Establish non-asymptotic error bounds using Wasserstein-1 distance.
result Non-asymptotic error bounds for discretization errors in distributional computational graphs.
Quantum computing promises new financial modeling.
problem Traditional financial modeling limitations.
method Overview of quantum computing applications in finance.
result Quantum computing can enhance financial modeling.
We look into computational aspects of two classical knot invariants. We look for ways of simplifying the computation of the coloring invariant and of the Alexander module. We support our ideas with explicit computations on pretzel knots.
Quantum computing techniques improve graph analysis and community detection.
problem Analyzing large graphs efficiently and accurately.
method Used quantum annealing and quantum gate computers for community detection and regularity checking.
result Demonstrated the effectiveness of quantum computing in solving complex graph problems.
New method speeds up knot computations in 3D.
problem Computational complexity in knot theory.
method 3D representation of knots for faster computation.
result Savings in computational complexity for knot invariants.
PALMS reconstructs large-scale networks efficiently with parallel computing.
problem Reconstructing large-scale latent networks from observed dynamics is computationally challenging.
method PALMS (Parallel Adaptive Lasso with Multi-directional Signals) framework for distributed network reconstruction.
result PALMS substantially reduces computational complexity and storage requirements.
Adaptive compute allocation improves model performance by prioritizing harder queries.
problem Inefficiency in allocating test-time compute uniformly across all queries.
method Formulated as a bandit learning problem, proposed adaptive algorithms that estimate query difficulty and allocate compute accordingly.
result Achieved up to 15.29% relative performance improvement on various benchmarks.
Optimizes K inner simulations for least-square Monte Carlo to reduce computational cost.
problem Computing conditional expectation E[f (Y)|X] with limited samples.
method Determines optimal number of Y samples (K) for given computational budget.
result Computational gain is maximized when sampling Y given X is inexpensive.