Everything is sole-authored unless listed otherwise. Preprints carry Zenodo DOIs; code links go to the repository that reproduces the reported numbers. Also on ORCID and ResearchGate.

Under review

  1. Persian Pronunciation Variation: Predictability and Consequences for Forced Alignment A. Ghanipour Amirhandeh, H. Asadi Submitted to ICPhS 2027Under review

    Six pronunciation lexicons (canonical, oracle, FST / logistic regression / BiLSTM phoneme-to-phone transducers) evaluated with Montreal Forced Aligner against manual boundaries on a 24-speaker corpus.

    Code
  2. Spectral Wasserstein Gradient Preconditioning for Non-Convex Frequency Optimization in Raw-Waveform Neural Networks A. Ghanipour Amirhandeh Under review, IEEE Transactions on Audio, Speech and Language ProcessingUnder review
  3. Adversarial Disentanglement of Speaker Identity and L2 Phonological Errors via Raw-Waveform Architectures A. Ghanipour Amirhandeh Submitted to NLP4CALL 2026Under review

Accepted

  1. Open Baselines and Controls for Conclusion Extraction on PARHAF (PARTAGES CU3) A. Ghanipour Amirhandeh Building Commons for Clinical LLMs International Conference, 2026Accepted

    First open baselines for a French clinical corpus, with degenerate and positional controls, paired bootstrap comparisons and an audit of the gold annotations.

    Code Results

Preprints and reports

  1. rPCEN: Scale-Equivariant Rapid-Release Per-Channel Energy Normalization via Parallel Adjoint Block Scans A. Ghanipour Amirhandeh Preprint, 2026Preprint DOI
  2. Monotonic Analytic Wavelet Front-Ends with Dynamic Non-Stationary Adaptation for Raw-Waveform Speaker Recognition A. Ghanipour Amirhandeh Preprint, 2026Preprint DOI
  3. Self-Supervised Speech Representations for Low-Resource and On-Device Streaming ASR: Architectural Paradigms, Compression Bottlenecks, and Open Challenges A. Ghanipour Amirhandeh Technical report, 2026Preprint DOI

In preparation

  1. Evaluating the Contribution of Speech Rhythm Measures to MFCC-Based Speaker Recognition H. Asadi, A. Ghanipour Amirhandeh In preparation, targeting Interspeech 2027In preparation Code
  2. Discrete RVQ Audio Watermarking via Hierarchical Residual Injection A. Ghanipour Amirhandeh Manuscript in preparation, targeting IEEE TASLPIn preparation

Thesis

  1. Speaker Recognition in Persian Spontaneous Speech using Deep Learning A. Ghanipour Amirhandeh M.A. thesis, University of Isfahan, 2026 · defended May 2026, summa cum laudeThesis Summary