Adjusting precision and topology parameters for neural network training based on a performance metric
US-12585926-B2 · Mar 24, 2026 · US
This page is not indexed by search engines while we improve data quality.
Lo Daniel patents
Lo Daniel is listed as an inventor on 66 patents in our database. Major assignees and classification codes are summarized below.
| Metric | Value |
|---|---|
| Inventor | Lo Daniel |
| Total patents | 66 |
| First publication | May 17, 2018 |
| Latest publication | Mar 24, 2026 |
Notable patents
Publications ranked by popularity score, then publication date.
Dithered quantization of parameters during training with a machine learning tool
US-12566962-B2 · Mar 3, 2026 · US
Compression and storage of neural network activations for backpropagation
US-2026017520-A1 · Jan 15, 2026 · US
Residual quantization for neural networks
US-12518151-B2 · Jan 6, 2026 · US
Neural network activation compression with narrow block floating-point
US-12443848-B2 · Oct 14, 2025 · US
Parallelized decoding of variable-length prefix codes
US-12438557-B2 · Oct 7, 2025 · US
Neural network activation compression with non-uniform mantissas
US-12277502-B2 · Apr 15, 2025 · US
Adjusting activation compression for neural network training
US-2025061320-A1 · Feb 20, 2025 · US
Adjusting activation compression for neural network training
US-12165038-B2 · Dec 10, 2024 · US
Neural network activation compression with non-uniform mantissas
US-12067495-B2 · Aug 20, 2024 · US
Recent patents
Latest publications not already listed above.
High-performance microcoded text parser
US-2024273287-A1 · Aug 15, 2024 · US
Neural network activation compression with outlier block floating-point
US-12045724-B2 · Jul 23, 2024 · US
High-performance microcoded text parser
US-11989508-B2 · May 21, 2024 · US
Neural network activation compression with non-uniform mantissas
US-2024152758-A1 · May 9, 2024 · US
High-performance table-based state machine
US-11863182-B2 · Jan 2, 2024 · US
Neural network training with decreased memory consumption and processor utilization
US-11853897-B2 · Dec 26, 2023 · US
Parallelized decoding of variable-length prefix codes
US-2023403028-A1 · Dec 14, 2023 · US
Training neural networks using mixed precision computations
US-11741362-B2 · Aug 29, 2023 · US
Training neural network accelerators using mixed precision data formats
US-2023267319-A1 · Aug 24, 2023 · US
Residual quantization for neural networks
US-2023196085-A1 · Jun 22, 2023 · US
Training neural network accelerators using mixed precision data formats
US-11676003-B2 · Jun 13, 2023 · US
Flow for quantized neural networks
US-11645493-B2 · May 9, 2023 · US
Neural network activation compression with non-uniform mantissas
US-2023140185-A1 · May 4, 2023 · US
Neural Network Training With Decreased Memory Consumption And Processor Utilization
US-2023110219-A1 · Apr 13, 2023 · US
Residual quantization for neural networks
US-11586883-B2 · Feb 21, 2023 · US
Outlier quantization for training and inference
US-11574239-B2 · Feb 7, 2023 · US
Neural network layer processing with normalization and transformation of data
US-11562201-B2 · Jan 24, 2023 · US
Neural network activation compression with non-uniform mantissas
US-11562247-B2 · Jan 24, 2023 · US
Deriving a concordant software neural network layer from a quantized firmware neural network layer
US-11556764-B2 · Jan 17, 2023 · US
Neural network layer processing with scaled quantization
US-11544521-B2 · Jan 3, 2023 · US
Major assignees
Companies most often associated with this inventor's publications.
| Assignee | Patents |
|---|---|
| Microsoft Technology Licensing Llc | 64 |
| Univ Cornell | 2 |
Top CPC classifications
Most common classification codes across this inventor's patents.
| CPC | Patents |
|---|---|
| G06N3/063 | 46 |
| G06N3/0495 | 36 |
| G06N3/084 | 34 |
| G06N3/0464 | 26 |
| G06F17/16 | 23 |