More Like this

1. Towards Vector Optimization on Low-Dimensional Vector Symbolic Architecture

   Duan, Shijin; Liu, Yejia; Liu, Gaowen; Kompella, Ramana Rao; Ren, Shaolei; Xu, Xiaolin (March 2025, The Conference on Parsimony and Learning (CPAL))

   Vector Symbolic Architecture (VSA) is emerging in machine learning due to its efficiency, but they are hindered by issues of hyperdimensionality and accuracy. As a promising mitigation, the Low-Dimensional Computing (LDC) method significantly reduces the vector dimension by 100 times while maintaining accuracy, by employing a gradient-based optimization. Despite its potential, LDC optimization for VSA is still underexplored. Our investigation into vector updates underscores the importance of stable, adaptive dynamics in LDC training. We also reveal the overlooked yet critical roles of batch normalization (BN) and knowledge distillation (KD) in standard approaches. Besides the accuracy boost, BN does not add computational overhead during inference, and KD significantly enhances inference confidence. Through extensive experiments and ablation studies across multiple benchmarks, we provide a thorough evaluation of our approach and extend the interpretability of binary neural network optimization similar to LDC, previously unaddressed in BNN literature. 

   more » « less
2. Mode Vector Modulation: Extending Stokes Vector Modulation to Higher Dimensions

   https://doi.org/10.1109/JLT.2023.3333801  

   Kwapisz, Jaroslaw; Roudas, Ioannis; Fink, Eric (March 2024, Journal of Lightwave Technology)
3. Vector autoregression and envelope model: Vector autoregression and envelope model

   https://doi.org/10.1002/sta4.203  

   Wang, Lei; Ding, Shanshan (January 2018, Stat)

   null (Ed.)
4. Mode Vector Modulation

   Roudas, Ioannis Kwapisz (September 2021, European Conference on Optical Communication)

   null (Ed.)

   We propose a new energy-efficient, short-haul, multidimensional modulation using spatial degrees of freedom in SDM fibers to create well-separated points in the generalized Stokes space. We study the transceiver architecture, geometric constellation shaping, bit-to-symbol mapping, and the performance of the optically-preamplified direct-detection receiver. 

   more » « less
5. Universal Vector Commitments

   https://doi.org/10.1007/978-3-031-71070-4_8  

   Acharya, Ojaswi; Baldimtsi, Foteini; Gordon, Samuel Dov; McVicker, Daniel; Yadav, Aayush (January 2024, Springer Nature Switzerland)