Conformer (neural network)
The Conformer is a neural network architecture that puts a convolution module and a self-attention module in the same block. It was introduced for automatic speech recognition by researchers at Google in 2020, in a paper titled "Conformer: Convolution-augmented Transformer for Speech Recognition".[1] The name is a contraction of convolution and transformer. Its encoder has since been used on other one-dimensional signals, among them nanopore reads, electroencephalograms and, in the case of Molecular Reality Corporation, the current through a single nanopore.
Background
Through self-attention, a transformer relates every position in a sequence to every other and so captures global context. A convolutional network looks at a small neighborhood at a time and captures local patterns well and cheaply. The authors say that transformers are good at "content-based global interactions" while convolutions "exploit local features effectively", and set out to combine the two "in a parameter-efficient way".[2] Their conclusion is that including the convolution modules is critical to the performance of the model.
Architecture
A Conformer block contains four modules in a fixed order. A feed-forward module comes first and a second feed-forward module comes last, each with its output added at half weight, a sandwich borrowed from an earlier design called Macaron-Net. Between them sit a multi-headed self-attention module that uses the relative sinusoidal positional encoding of Transformer-XL, and a convolution module. The convolution module has a pointwise convolution that doubles the channels, a gated linear unit, a depthwise one-dimensional convolution, batch normalization, a Swish activation and a second pointwise convolution. A layer normalization follows the block.[2]
The input is a spectrogram, with 80 filterbank channels computed over a 25 ms window every 10 ms. A convolutional layer subsamples it before the blocks. Three sizes were trained on the 970 hours of LibriSpeech, a corpus of read audiobooks.
| Model | Parameters | Layers | Encoder width | Attention heads | Convolution kernel |
|---|---|---|---|---|---|
| Small | 10.3 million | 16 | 144 | 4 | 32 |
| Medium | 30.7 million | 16 | 256 | 4 | 32 |
| Large | 118.8 million | 17 | 512 | 8 | 32 |
The large model reached word error rates of 2.1% on the standard test set and 4.3% on the harder "test-other" set without a language model, and 1.9% and 3.9% with an external one, which the paper calls a new state of the art, and a 15% relative improvement on test-other with an external language model over the best previous model, a Transformer Transducer. The small model reached 2.7% and 6.3%.[2]
Other signals
Nanopore sequencing was an early destination. A basecaller of 2020 built on the convolution-augmented transformer decoded the raw current of an Oxford Nanopore device into bases, and a basecaller for nanopore RNA in 2024 combined a transformer with gated convolution.[3][4] For EEG, a network that places convolutional layers ahead of a transformer was published under the name EEG Conformer.[5]
A speech model takes one feature vector every 10 ms. The company's live stream records a value every 5 µs, which is 2,000 times faster, and at 100 kS/s an event of 185 µs is about 18 samples long.
At Molecular Reality
In April 2026 the company trained a Conformer of three blocks on barcode events from a public set of DNA barcodes read by a solid-state nanopore, and compared it with a retrained QuipuNet CNN on a split that held out whole experiments. The Conformer scored 85.4% against 81.0% for the CNN, a result that is unpublished.[6] Its description reads "experiments replicating existing nanopore/speech applications of novel AI methods", and one of its steps reruns, in simplified form, the speech experiment of the Conformer paper.[6] Company records call the model the first Conformer the company knows of on single-molecule sensing traces, a qualifier that matters, since basecallers had used the design before.[7]
In the company's plans the Conformer reads the current, a Perceiver fuses current, voltage and flow-cell video, and a joint embedding predictive architecture supplies a training objective that needs no labels, as laid out in The Nanopore World Model.[7][8] Its investor copy puts the case for the whole program in one line: "The machine learning that cracked speech now works on ionic current."[7]
See also
References
- ^ Gulati, A.; et al. (2020). "Conformer: Convolution-augmented Transformer for Speech Recognition". pp. 5036-5040. doi:10.21437/Interspeech.2020-3015.
- ^ a b c Gulati, A.; et al. (2020). "Conformer: Convolution-augmented Transformer for Speech Recognition". arXiv. doi:10.48550/arXiv.2005.08100.
- ^ Lv, X.; et al. (2020). "An end-to-end Oxford Nanopore basecaller using convolution-augmented transformer". doi:10.1109/BIBM49941.2020.9313290.
- ^ Li, Q.; et al. (2024). "GCRTcall: a transformer based basecaller for nanopore RNA sequencing enhanced by gated convolution and relative position embedding via joint loss training". Frontiers in Genetics 15: 1443532. doi:10.3389/fgene.2024.1443532.
- ^ Song, Y.; et al. (2023). "EEG Conformer: convolutional transformer for EEG decoding and visualization". IEEE Transactions on Neural Systems and Rehabilitation Engineering 31. doi:10.1109/TNSRE.2022.3230250.
- ^ a b "ai-experiments". GitHub. April 2026.
- ^ a b c "Company records". Molecular Reality Corporation. October 2026.
- ^ "Keltar's Critique v2: The Nanopore World Model (Hardened)". Molecular Reality Corporation. June 2026.