New answers tagged reference-request
0
votes
Accepted
How can machine learning be used to create synthesized sounds of world instruments?
DeepMind's WaveNet initially developed for speech synthesis based directly upon raw audio waveforms model has been adapted for generating realistic audio including musical instruments.
WaveNet's ...
- 5,387
1
vote
Which factors do distinguish between human and LLMs written text at the moment?
OpenAI decided to discontinue its GPT text detection tool mainly due to its low accuracy. This high rate of errors suggests that distinguishing between human and LLM-written content remains a ...
- 5,387
1
vote
Are there any metrics suitable for evaluating conditional GANs in the context of medical imaging?
Most metrics for evaluating generative models, including IS and FID, are indeed based on pre-trained feature extractors. Regarding the data where these features extractors are trained, I agree with ...
- 153
2
votes
Are there any metrics suitable for evaluating conditional GANs in the context of medical imaging?
Indeed instead of relying on pre-trained models from natural image datasets such as ImageNet or COCO-2014, you should train a classification model as feature extractor specifically on your concerned ...
- 5,387
0
votes
Lower Bound for LLM size?
Kaplan et al. (2020) "Scaling Laws for Neural Language Models" established fundamental relationships between model size, compute, performance and showed power-law relationships in model ...
- 5,387
1
vote
Accepted
Are there any attempts to make translation models support ancient/extinct languages?
I am not aware of any attempts at building machine translation for extinct languages specifically, but I would see this as a sub-field of machine translation for low resource languages. The techniques ...
Top 50 recent answers are included