Binarized Neural Networks: Training Deep Neural Networks with Weights and Activations Constrained to +1 or -1 (2016 · 2,143 citations)
This work tackles the cost of running deep networks on low-power hardware. Modern DNNs depend on energy-hungry GPUs and full-precision arithmetic. The authors propose Binarized Neural Networks (BNNs), in which weights and hidden activations take only the values +1 or −1. These binary values are used both at inference…
Quantized Neural Networks: Training Neural Networks with Low Precision Weights and Activations (2016 · 1,610 citations)
The paper tackles the high cost of running deep networks, whose many multiply-accumulate operations make them slow and power-hungry on ordinary hardware and low-power devices.
Llama-Nemotron: Efficient Reasoning Models (2025 · 76 citations)
Reasoning-oriented language models write long chains of thought, so inference cost increasingly limits how useful they are.

Co-authors here

6 co-authors with a page, the most papers together first.

Cited by

Papers of our library cite their works 326 times (of the 149,439 papers whose bibliographies we read); 48 citing papers with a page here →.

Updated 2 November 2026