# Automatic Speech Recognition on HKUST

Source: [Automatic Speech Recognition on HKUST](<https://www.sota2.com/research/sota/automatic-speech-recognition-on-hkust>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-02-26T06:04:51.501Z

Evaluation of Automatic Speech Recognition performance on the HKUST dataset, typically measured using Character Error Rate (CER) and its components (substitutions, deletions, insertions).

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

| Scope | Recorded value |
| --- | --- |
| Task | Automatic Speech Recognition |
| Dataset (including any stated split) | HKUST |
| Metric | Sub Error Rate |
| Unit | PERCENT |
| Direction | Lower is better |
| Results loaded | 18 of 18 |
| Coverage | All results currently returned for this benchmark |

Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol.

## Metrics

- [Sub Error Rate](<https://www.sota2.com/research/sota/automatic-speech-recognition-on-hkust.md?metric=Sub%20Error%20Rate>): Lower is better; unit: PERCENT.

- [Del Error Rate](<https://www.sota2.com/research/sota/automatic-speech-recognition-on-hkust.md?metric=Del%20Error%20Rate>): Lower is better; unit: PERCENT.

- [Ins Error Rate](<https://www.sota2.com/research/sota/automatic-speech-recognition-on-hkust.md?metric=Ins%20Error%20Rate>): Lower is better; unit: PERCENT.

- [Character Error Rate (CER)](<https://www.sota2.com/research/sota/automatic-speech-recognition-on-hkust.md?metric=Character%20Error%20Rate%20(CER)>): Lower is better; unit: PERCENT.

## Recorded results

| Method | Sub Error Rate | Result date | Report paper |
| --- | --- | --- | --- |
| UMA + self-condition | 13.7 | Not recorded | 68fa72911e209131cea57245 |
| Hybrid CTC/Attention + beam search | 14.1 | Not recorded | 68fa72911e209131cea57245 |
| UMA | 14.1 | Not recorded | 68fa72911e209131cea57245 |
| UMA + self-condition | 14.4 | Not recorded | 68fa72911e209131cea57245 |
| Self-conditioned CTC | 14.9 | Not recorded | 68fa72911e209131cea57245 |
| Hybrid CTC/Attention | 15.2 | Not recorded | 68fa72911e209131cea57245 |
| UMA | 15.6 | Not recorded | 68fa72911e209131cea57245 |
| Hybrid CTC/Attention + beam search | 15.7 | Not recorded | 68fa72911e209131cea57245 |
| UMA + self-condition | 15.8 | Not recorded | 68fa72911e209131cea57245 |
| Hybrid CTC/Attention + beam search | 15.9 | Not recorded | 68fa72911e209131cea57245 |
| UMA | 15.9 | Not recorded | 68fa72911e209131cea57245 |
| CTC | 16 | Not recorded | 68fa72911e209131cea57245 |
| Self-conditioned CTC | 16.3 | Not recorded | 68fa72911e209131cea57245 |
| Hybrid CTC/Attention | 16.9 | Not recorded | 68fa72911e209131cea57245 |
| CTC | 17.3 | Not recorded | 68fa72911e209131cea57245 |
| Hybrid CTC/Attention | 18 | Not recorded | 68fa72911e209131cea57245 |
| Self-conditioned CTC | 18.3 | Not recorded | 68fa72911e209131cea57245 |
| CTC | 18.4 | Not recorded | 68fa72911e209131cea57245 |

### UMA + self-condition

Sub Error Rate: 13.7. Result ID: 69925d3a3a6c3148490e747e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72911e209131cea57245.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72911e209131cea57245.md>)

[Implementation](<https://github.com/Audio-WestlakeU/UMA-ASR>)

Recorded evaluation configuration: {"Encoder":"E-Branchformer","Category":"NAR"}. This record does not assert independent reproduction.

### Hybrid CTC/Attention + beam search

Sub Error Rate: 14.1. Result ID: 69925d3a3a6c3148490e747a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72911e209131cea57245.md>)

Recorded evaluation configuration: {"Encoder":"E-Branchformer","Category":"AR","Decoding":"Beam Search"}. This record does not assert independent reproduction.

### UMA

Sub Error Rate: 14.1. Result ID: 69925d3a3a6c3148490e747d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72911e209131cea57245.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72911e209131cea57245.md>)

[Implementation](<https://github.com/Audio-WestlakeU/UMA-ASR>)

Recorded evaluation configuration: {"Encoder":"E-Branchformer","Category":"NAR"}. This record does not assert independent reproduction.

### UMA + self-condition

Sub Error Rate: 14.4. Result ID: 69925d3a3a6c3148490e7478.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72911e209131cea57245.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72911e209131cea57245.md>)

[Implementation](<https://github.com/Audio-WestlakeU/UMA-ASR>)

Recorded evaluation configuration: {"Encoder":"Conformer","Category":"NAR"}. This record does not assert independent reproduction.

### Self-conditioned CTC

Sub Error Rate: 14.9. Result ID: 69925d3a3a6c3148490e747c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72911e209131cea57245.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6af91e209131cea2b75a.md>)

Recorded evaluation configuration: {"Encoder":"E-Branchformer","Category":"NAR"}. This record does not assert independent reproduction.

### Hybrid CTC/Attention

Sub Error Rate: 15.2. Result ID: 69925d3a3a6c3148490e7479.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72911e209131cea57245.md>)

Recorded evaluation configuration: {"Encoder":"E-Branchformer","Category":"AR"}. This record does not assert independent reproduction.

### UMA

Sub Error Rate: 15.6. Result ID: 69925d3a3a6c3148490e7477.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72911e209131cea57245.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72911e209131cea57245.md>)

[Implementation](<https://github.com/Audio-WestlakeU/UMA-ASR>)

Recorded evaluation configuration: {"Encoder":"Conformer","Category":"NAR"}. This record does not assert independent reproduction.

### Hybrid CTC/Attention + beam search

Sub Error Rate: 15.7. Result ID: 69925d3a3a6c3148490e7474.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72911e209131cea57245.md>)

Recorded evaluation configuration: {"Encoder":"Conformer","Category":"AR","Decoding":"Beam Search"}. This record does not assert independent reproduction.

### UMA + self-condition

Sub Error Rate: 15.8. Result ID: 69925d3a3a6c3148490e7472.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72911e209131cea57245.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72911e209131cea57245.md>)

[Implementation](<https://github.com/Audio-WestlakeU/UMA-ASR>)

Recorded evaluation configuration: {"Encoder":"Transformer","Category":"NAR"}. This record does not assert independent reproduction.

### Hybrid CTC/Attention + beam search

Sub Error Rate: 15.9. Result ID: 69925d3a3a6c3148490e746e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72911e209131cea57245.md>)

Recorded evaluation configuration: {"Encoder":"Transformer","Category":"AR","Decoding":"Beam Search"}. This record does not assert independent reproduction.

### UMA

Sub Error Rate: 15.9. Result ID: 69925d3a3a6c3148490e7471.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72911e209131cea57245.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72911e209131cea57245.md>)

[Implementation](<https://github.com/Audio-WestlakeU/UMA-ASR>)

Recorded evaluation configuration: {"Encoder":"Transformer","Category":"NAR"}. This record does not assert independent reproduction.

### CTC

Sub Error Rate: 16. Result ID: 69925d3a3a6c3148490e747b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72911e209131cea57245.md>)

Recorded evaluation configuration: {"Encoder":"E-Branchformer","Category":"NAR"}. This record does not assert independent reproduction.

### Self-conditioned CTC

Sub Error Rate: 16.3. Result ID: 69925d3a3a6c3148490e7476.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72911e209131cea57245.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6af91e209131cea2b75a.md>)

Recorded evaluation configuration: {"Encoder":"Conformer","Category":"NAR"}. This record does not assert independent reproduction.

### Hybrid CTC/Attention

Sub Error Rate: 16.9. Result ID: 69925d3a3a6c3148490e7473.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72911e209131cea57245.md>)

Recorded evaluation configuration: {"Encoder":"Conformer","Category":"AR"}. This record does not assert independent reproduction.

### CTC

Sub Error Rate: 17.3. Result ID: 69925d3a3a6c3148490e7475.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72911e209131cea57245.md>)

Recorded evaluation configuration: {"Encoder":"Conformer","Category":"NAR"}. This record does not assert independent reproduction.

### Hybrid CTC/Attention

Sub Error Rate: 18. Result ID: 69925d3a3a6c3148490e746d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72911e209131cea57245.md>)

Recorded evaluation configuration: {"Encoder":"Transformer","Category":"AR"}. This record does not assert independent reproduction.

### Self-conditioned CTC

Sub Error Rate: 18.3. Result ID: 69925d3a3a6c3148490e7470.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72911e209131cea57245.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6af91e209131cea2b75a.md>)

Recorded evaluation configuration: {"Encoder":"Transformer","Category":"NAR"}. This record does not assert independent reproduction.

### CTC

Sub Error Rate: 18.4. Result ID: 69925d3a3a6c3148490e746f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72911e209131cea57245.md>)

Recorded evaluation configuration: {"Encoder":"Transformer","Category":"NAR"}. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.
