[HN Gopher] Jargonic: Industry-Tunable ASR Model
       ___________________________________________________________________
        
       Jargonic: Industry-Tunable ASR Model
        
       Author : agold97
       Score  : 53 points
       Date   : 2025-04-01 07:35 UTC (15 hours ago)
        
 (HTM) web link (aiola.ai)
 (TXT) w3m dump (aiola.ai)
        
       | four_fifths wrote:
       | so if i understand this correctly -- you want the speech
       | recognition model to identify a vocabulary of specific terms that
       | it wasn't trained on. instead of fine-tuning with training data
       | that includes the new vocabulary, you input the full vocabulary
       | at test time as a list of words and the model is able to generate
       | transcripts that include words from the vocabulary.
       | 
       | seems like it could be very useful but it really comes down to
       | the specifics.
       | 
       | you can prompt whisper with context -- how does this compare?
       | 
       | how large of a vocabulary can it work with? if it's a few dozen
       | words it's only gonna help for niche use cases. if it can handle
       | 100s-1000s with good performance that could completely replace
       | fine-tuning for many uses
        
         | GavCo wrote:
         | I was wondering the same and found these related papers:
         | 
         | https://arxiv.org/pdf/2309.08561
         | https://arxiv.org/pdf/2406.02649
         | 
         | I haven't really dug in yet but from a quick skim, it looks
         | promising. They show a big improvement over Whisper on a
         | medical dataset (F1 increased from 80.5% to 96.58%).
         | 
         | The inference time for the keyword detection is about 10ms. If
         | it scales linearly with additional keywords you could
         | potentially scale to hundreds or thousands of keywords but it
         | really depends on how sensitive you are to latency. For real-
         | time with large vocabularies my guess is you might still want
         | to fine-tune.
        
         | agold97 wrote:
         | yeah -- sounds about right. retraining the whole model just to
         | add one jargon-y term isn't super efficient. this approach lets
         | you plug in a vocab list at runtime instead, which feels a lot
         | more scalable.
        
       | gronky_ wrote:
       | I just tried the demo on the homepage and I don't know what kind
       | of sorcery this is but it's blowing my mind.
       | 
       | I input a bunch of completely made up words (Quastral Syncing,
       | Zarnix Meshing, HIBAX, Bilxer) and used them in a sentence and
       | the model zero-shotted perfect speech recognition!
       | 
       | It's so counterintuitive for me that this would work. I would
       | have bet that you have to provide at least one audio sample in
       | order for the model to recognize a word it was never trained on.
       | 
       | Providing it to the model in text modality and it being able to
       | recognize it in the audio modality must be an emergent property.
        
       | htrp wrote:
       | perhaps it's using openai advanced voice or another tts to create
       | waveforms for comparison?
        
       | FloatArtifact wrote:
       | How does this keyword spotting compare versus grammar or intent
       | approach for speech recognition commands with dictation?
       | 
       | How does keyword spotting handle complex phrases as commands?
        
       | suchire wrote:
       | Is their WER graph just completely made up? It's comically bad
        
       ___________________________________________________________________
       (page generated 2025-04-01 23:02 UTC)