Global ETD Search

Search theses and dissertations gathered from participating repositories worldwide. Every result links back to the library that holds it. No account is needed.

Results

Showing 1 to 14 of 14 for “"Finite Precision"”.

  1. Finite precision deep learning with theoretical guarantees

    … popular way of doing so is via implementation in finite precision. While ad-hoc trial and error techniques in finite precision deep learning abound, theoretical guarantees on network accuracy are elusive. The work presented in this dissertation builds a theoretical framework for the implementation …

    uiuc Repository record for Finite precision deep learning with theoretical guarantees (opens in a new tab)

  2. Control and diagnosis of real-time systems under finite-precision measurement of time

    … dense-time DESs in which the digital-clocks of finite-precision are employed to observe the event occurrence times, thereby relaxing the assumption of the prior works that time can be measured precisely. In our setting, the passing of time is measured using the number of ticks generated by a …

    iastate Repository record for Control and diagnosis of real-time systems under finite-precision measurement of time (opens in a new tab)

  3. On the computational power of RNNs

    … power of such systems? We prove that finite precision RNNs with one hidden layer and ReLU activation and finite precision GRUs are exactly as computationally powerful as deterministic nice automata. Allowing arbitrary precision, we prove that RNNs with one hidden layer and ReLU …

    mit Repository record for On the computational power of RNNs (opens in a new tab)

  4. Wavelets and multirate filter banks : theory, structure, design, and applications

    … is structurally enforced, even under finite precision arithmetic and quantization of lifting coefficients. With lifting, optimal low-complexity integer wavelet transforms can thus be designed using a simple and fast algorithm that incorporates prescribed limits on hardware operations …

    mit Repository record for Wavelets and multirate filter banks : theory, structure, design, and applications (opens in a new tab)

  5. Topological entropy bounds for switched linear systems with lie structure

    … needed to describe the behavior of a system with finite precision on finite time horizons, in the limit. After working out entropy computations in detail for the scalar switched case, we review the linear time-invariant nonscalar case, and extend to the nonscalar switched case. We assume some …

    uiuc Repository record for Topological entropy bounds for switched linear systems with lie structure (opens in a new tab)

  6. The analysis and synthesis of efficient algorithm-based error detection schemes for hypercube multiprocessors

    … error detection schemes in the presence of finite precision arithmetic which affects our system-level encodings.

    uiuc Repository record for The analysis and synthesis of efficient algorithm-based error detection schemes for hypercube multiprocessors (opens in a new tab)

  7. Discrete Quantum Theories and Computing

    … mathematical ways to limit measurement precision to computable contexts that do not require incomputable real numbers. Our goal is to build a more philosophically consistent models by investigating discrete quantum comput- ing using finite number systems, and, alternatively, by …

    iu Repository record for Discrete Quantum Theories and Computing (opens in a new tab)

  8. Difference-operator-based models in statistical signal processing

    … numerical performance when implemented using finite-precision arithmetic. Unlike the situation for shift-operator models, estimating the parameters of a difference-operator-based model involves solving a non-Toeplitz system of linear equations. We derive an algorithm for efficiently estimating …

    uiuc Repository record for Difference-operator-based models in statistical signal processing (opens in a new tab)

  9. On the Foundations of Computation and Sampling for Reconstruction and Approximation

    … of learning from point samples with finite precision. In the future we aim to extend this work to give a guideline on the stable usage of deep learning with neural networks.

    cambridge Repository record for On the Foundations of Computation and Sampling for Reconstruction and Approximation (opens in a new tab)

  10. Implementation and evaluation of echo cancellation algorithms

    … system, such as double-talk detection, finite precision effects, nonlinear processing, and howling detection and control. The use of a second adaptive filter is proposed, to reduce near-end ambient noise. Simulation results indicate that this approach can reduce the ambient noise by …

    vt Repository record for Implementation and evaluation of echo cancellation algorithms (opens in a new tab)

  11. Signal processing in a loudspeaking telephone

    … the adptation which can be achieved using finite-precision arithmatic. This enabled the development of an architecture which distributed the normalisation required to achieve optimum rate of adaptation over the useful input range. Comparison of theory and measurement for the adaptive filter …

    aston Repository record for Signal processing in a loudspeaking telephone (opens in a new tab)

  12. DCD Algorithm: Architectures, FPGA Implementations and Applications

    … regularized RLS-DCD algorithm is simple for finite precision implementation and requires small chip resources.

    whiterose Repository record for DCD Algorithm: Architectures, FPGA Implementations and Applications (opens in a new tab)

  13. Algorithms and Algorithmic Barriers in High-Dimensional Statistics and Random Combinatorial Structures

    … and is widely believed to be true) both for the finite-precision arithmetic model and for the real-valued computational model. Our result is the first provable hardness result for a statistical physics model with random parameters that is based on standard complexity-theoretical assumptions. 3. …

    mit Repository record for Algorithms and Algorithmic Barriers in High-Dimensional Statistics and Random Combinatorial Structures (opens in a new tab)

  14. A digital interface for wireless networks

    … and by restricting the transmit signals to a finite alphabet. The superposition network, inspired by the Gaussian model, is a noiseless deterministic network, the inputs and outputs of the channels are discrete, and channel gains are signed integers. The capacity of a class of Gaussian relay …

    uiuc Repository record for A digital interface for wireless networks (opens in a new tab)