본문

서브메뉴

Learning Audio Foundation Models for Reasoning
Learning Audio Foundation Models for Reasoning
Learning Audio Foundation Models for Reasoning

상세정보

자료유형  
 학위논문 서양
최종처리일시  
20260202103610
ISBN  
9798290940250
DDC  
534
저자명  
Deshmukh, Soham.
서명/저자  
Learning Audio Foundation Models for Reasoning
발행사항  
[Sl] : Carnegie Mellon University, 2025
발행사항  
Ann Arbor : ProQuest Dissertations & Theses, 2025
형태사항  
248 p
주기사항  
Source: Dissertations Abstracts International, Volume: 87-02, Section: B.
주기사항  
Advisor: Raj, Bhiksha.
학위논문주기  
Thesis (Ph.D.)--Carnegie Mellon University, 2025.
초록/해제  
요약Sound understanding, a field of machine learning, has applications across healthcare, noise monitoring, bioacoustics, home security, and music. Traditionally, task-specific models were developed for each application, requiring data collection, annotation, model training, evaluation, and deployment. This iterative process, which involves designing architectures, datasets, and objectives, is repeated for each new domain, task, or application.To streamline this process, recent research has focused on pretraining large-scale datasets to learn general-purpose audio representations or embeddings. These pretrained models can then be adapted to new tasks by training linear or shallow classifiers on top of the learned embeddings. Pretrained models with fine-tuning consistently outperform task-specific models in downstream tasks, making them a cornerstone of modern sound understanding.Given the success of pretraining and scaling (both in compute and data), this thesis investigates two central questions: First, what are effective learning or pretraining methods to enable the scaling of audio models? Second, once pretrained, how can the new abilities be analyzed, evaluated, and subsequently improved? These questions are interdependent, with advancements in one area feeding into and enhancing the other. The resulting models, termed Audio Foundation Models (AFMs), are pretrained on large-scale data and develop novel capabilities for understanding and reasoning about sound. This thesis aims to devise such models capable of understanding and reasoning about audio.In the first part of this thesis, we introduce scalable learning methods to develop AFMs. Specifically, we explore audio-language pretraining to connect auditory perception with linguistic understanding. We propose three learning methods and subsequently models. First, we propose CLAP (Contrastive Audio-Language Pretraining), which learns multimodal audio-text representations through contrastive learning. CLAP excels in retrieval and indexing tasks and supports zero-shot classification, where test-time classes are user-defined. Second, we introduce audio-conditioned text generation and subsequently Pengi, an audio-language model that frames audio tasks as text-generation tasks. By combining an audio encoder with a language model, Pengi enables open-ended tasks such as audio captioning and question-answering while maintaining strong performance on close-ended tasks. Finally, we introduce Mellow, a small audio-language model designed to address Pengi's limitations in detailed reasoning and question-answering, thereby improving audio-grounded reasoning capabilities.These scalable learning methods enable training on millions of audio-text pairs, resulting in Audio Foundation Models with emergent capabilities. These models often demonstrate abilities they were not explicitly trained for, making them valuable for addressing new domains and tasks. However, discovering and benchmarking these capabilities is critical to ensuring their safe and effective use in real-world applications. In the second part of this thesis, we focus on the reasoning and analytical abilities of Audio Foundation Models. First, we demonstrate that AFMs can serve as reference-free audio quality estimators using antonym prompting strategies. Second, we define logical reasoning for AFMs and introduce an audio entailment task to measure deductive reasoning, identifying key limitations of existing models. To enhance reasoning performance, we propose the "caption-before-reason" method, which can be applied directly at test time. Finally, we explore comparative reasoning across multiple audios, introducing two new datasets with detailed explanations and proposing the ADIFF model, which improves reasoning over state-of-the-art AFMs.While pretraining on large datasets is effective, it is resource-intensive and does not fully address long-tail and out-of-distribution scenarios. To address these challenges, in the third and final part of this thesis, we propose methods to adapt AFM to new domains with minimal resources. These methods are categorized based on the required data and computational budget. First, we introduce test-time adaptation, which adapts models to new data at test time using unlabeled audio. Second, we show that few-shot learning in Pengi can improve performance with limited labeled examples. Finally, we propose a method that leverages CLAP's multimodal space to enable missing-modality training for text generation tasks.In summary, this thesis presents scalable learning methods, evaluation benchmarks, and key components for developing Audio Foundation Models capable of reasoning. These models have diverse applications, including media retrieval, creative audio generation, healthcare assistance, and predictive maintenance in industrial settings. Looking ahead, we believe this work is crucial in the inevitable rise of audio agents, enabling embodied AI systems to understand and reason about their environments using rich acoustic information.
일반주제명  
Acoustics
일반주제명  
Computer engineering
일반주제명  
Electrical engineering
키워드  
Audio foundation models
키워드  
Audio processing
키워드  
Audio-language models
키워드  
Audio-text learning
키워드  
Machine listening
키워드  
Multimodal learning
기타저자  
Carnegie Mellon University Electrical and Computer Engineering
기본자료저록  
Dissertations Abstracts International. 87-02B.
전자적 위치 및 접속  
로그인 후 원문을 볼 수 있습니다.

MARC

 008260126s2025        us                              c    eng  d
■001000017357864
■00520260202103610
■006m          o    d                
■007cr#unu||||||||
■020    ▼a9798290940250
■035    ▼a(MiAaPQ)AAI32043167
■040    ▼aMiAaPQ▼cMiAaPQ
■0820  ▼a534
■1001  ▼aDeshmukh,  Soham.
■24510▼aLearning  Audio  Foundation  Models  for  Reasoning
■260    ▼a[Sl]▼bCarnegie  Mellon  University▼c2025
■260  1▼aAnn  Arbor▼bProQuest  Dissertations  &  Theses▼c2025
■300    ▼a248  p
■500    ▼aSource:  Dissertations  Abstracts  International,  Volume:  87-02,  Section:  B.
■500    ▼aAdvisor:  Raj,  Bhiksha.
■5021  ▼aThesis  (Ph.D.)--Carnegie  Mellon  University,  2025.
■520    ▼aSound  understanding,  a  field  of  machine  learning,  has  applications  across  healthcare,  noise  monitoring,  bioacoustics,  home  security,  and  music.  Traditionally,  task-specific  models  were  developed  for  each  application,  requiring  data  collection,  annotation,  model  training,  evaluation,  and  deployment.  This  iterative  process,  which  involves  designing  architectures,  datasets,  and  objectives,  is  repeated  for  each  new  domain,  task,  or  application.To  streamline  this  process,  recent  research  has  focused  on  pretraining  large-scale  datasets  to  learn  general-purpose  audio  representations  or  embeddings.  These  pretrained  models  can  then  be  adapted  to  new  tasks  by  training  linear  or  shallow  classifiers  on  top  of  the  learned  embeddings.  Pretrained  models  with  fine-tuning  consistently  outperform  task-specific  models  in  downstream  tasks,  making  them  a  cornerstone  of  modern  sound  understanding.Given  the  success  of  pretraining  and  scaling  (both  in  compute  and  data),  this  thesis  investigates  two  central  questions:  First,  what  are  effective  learning  or  pretraining  methods  to  enable  the  scaling  of  audio  models?  Second,  once  pretrained,  how  can  the  new  abilities  be  analyzed,  evaluated,  and  subsequently  improved?  These  questions  are  interdependent,  with  advancements  in  one  area  feeding  into  and  enhancing  the  other.  The  resulting  models,  termed  Audio  Foundation  Models  (AFMs),  are  pretrained  on  large-scale  data  and  develop  novel  capabilities  for  understanding  and  reasoning  about  sound.  This  thesis  aims  to  devise  such  models  capable  of  understanding  and  reasoning  about  audio.In  the  first  part  of  this  thesis,  we  introduce  scalable  learning  methods  to  develop  AFMs.  Specifically,  we  explore  audio-language  pretraining  to  connect  auditory  perception  with  linguistic  understanding.  We  propose  three  learning  methods  and  subsequently  models.  First,  we  propose  CLAP  (Contrastive  Audio-Language  Pretraining),  which  learns  multimodal  audio-text  representations  through  contrastive  learning.  CLAP  excels  in  retrieval  and  indexing  tasks  and  supports  zero-shot  classification,  where  test-time  classes  are  user-defined.  Second,  we  introduce  audio-conditioned  text  generation  and  subsequently  Pengi,  an  audio-language  model  that  frames  audio  tasks  as  text-generation  tasks.  By  combining  an  audio  encoder  with  a  language  model,  Pengi  enables  open-ended  tasks  such  as  audio  captioning  and  question-answering  while  maintaining  strong  performance  on  close-ended  tasks.  Finally,  we  introduce  Mellow,  a  small  audio-language  model  designed  to  address  Pengi's  limitations  in  detailed  reasoning  and  question-answering,  thereby  improving  audio-grounded  reasoning  capabilities.These  scalable  learning  methods  enable  training  on  millions  of  audio-text  pairs,  resulting  in  Audio  Foundation  Models  with  emergent  capabilities.  These  models  often  demonstrate  abilities  they  were  not  explicitly  trained  for,  making  them  valuable  for  addressing  new  domains  and  tasks.  However,  discovering  and  benchmarking  these  capabilities  is  critical  to  ensuring  their  safe  and  effective  use  in  real-world  applications.  In  the  second  part  of  this  thesis,  we  focus  on  the  reasoning  and  analytical  abilities  of  Audio  Foundation  Models.  First,  we  demonstrate  that  AFMs  can  serve  as  reference-free  audio  quality  estimators  using  antonym  prompting  strategies.  Second,  we  define  logical  reasoning  for  AFMs  and  introduce  an  audio  entailment  task  to  measure  deductive  reasoning,  identifying  key  limitations  of  existing  models.  To  enhance  reasoning  performance,  we  propose  the  "caption-before-reason"  method,  which  can  be  applied  directly  at  test  time.  Finally,  we  explore  comparative  reasoning  across  multiple  audios,  introducing  two  new  datasets  with  detailed  explanations  and  proposing  the  ADIFF  model,  which  improves  reasoning  over  state-of-the-art  AFMs.While  pretraining  on  large  datasets  is  effective,  it  is  resource-intensive  and  does  not  fully  address  long-tail  and  out-of-distribution  scenarios.  To  address  these  challenges,  in  the  third  and  final  part  of  this  thesis,  we  propose  methods  to  adapt  AFM  to  new  domains  with  minimal  resources.  These  methods  are  categorized  based  on  the  required  data  and  computational  budget.  First,  we  introduce  test-time  adaptation,  which  adapts  models  to  new  data  at  test  time  using  unlabeled  audio.  Second,  we  show  that  few-shot  learning  in  Pengi  can  improve  performance  with  limited  labeled  examples.  Finally,  we  propose  a  method  that  leverages  CLAP's  multimodal  space  to  enable  missing-modality  training  for  text  generation  tasks.In  summary,  this  thesis  presents  scalable  learning  methods,  evaluation  benchmarks,  and  key  components  for  developing  Audio  Foundation  Models  capable  of  reasoning.  These  models  have  diverse  applications,  including  media  retrieval,  creative  audio  generation,  healthcare  assistance,  and  predictive  maintenance  in  industrial  settings.  Looking  ahead,  we  believe  this  work  is  crucial  in  the  inevitable  rise  of  audio  agents,  enabling  embodied  AI  systems  to  understand  and  reason  about  their  environments  using  rich  acoustic  information.
■590    ▼aSchool  code:  0041.
■650  4▼aAcoustics
■650  4▼aComputer  engineering
■650  4▼aElectrical  engineering
■653    ▼aAudio  foundation  models
■653    ▼aAudio  processing
■653    ▼aAudio-language  models
■653    ▼aAudio-text  learning
■653    ▼aMachine  listening
■653    ▼aMultimodal  learning
■690    ▼a0800
■690    ▼a0986
■690    ▼a0544
■690    ▼a0464
■71020▼aCarnegie  Mellon  University▼bElectrical  and  Computer  Engineering.
■7730  ▼tDissertations  Abstracts  International▼g87-02B.
■790    ▼a0041
■791    ▼aPh.D.
■792    ▼a2025
■793    ▼aEnglish
■85640▼uhttp://www.riss.kr/pdu/ddodLink.do?id=T17357864▼nKERIS▼z이  자료의  원문은  한국교육학술정보원에서  제공합니다.

미리보기

내보내기

chatGPT토론

Ai 추천 관련 도서


    신착도서 더보기
    최근 3년간 통계입니다.

    소장정보

    • 예약
    • 소재불명신고
    • 나의폴더
    • 우선정리요청
    • 비도서대출신청
    • 야간 도서대출신청
    소장자료
    등록번호 청구기호 소장처 대출가능여부 대출정보
    TF15029 전자도서 대출가능 마이폴더 부재도서신고 비도서대출신청 야간 도서대출신청

    * 대출중인 자료에 한하여 예약이 가능합니다. 예약을 원하시면 예약버튼을 클릭하십시오.

    해당 도서를 다른 이용자가 함께 대출한 도서

    관련 인기도서

    로그인 후 이용 가능합니다.