Decoding Communication in Nonhuman Species IV
2025-07-30..2025-07-31
Berkeley, United States
who it's for
bioacoustics researchers, machine learning engineers for audio, computational ethologists, conservation technologists, field biologists, computational linguists interested in animal communication, data scientists for ecology
topics
sperm whale coda modeling; audio-to-audio coda translation; pre-trained acoustic embeddings; self-supervised spectrogram transformers; multimodal audio-visual alignment; dolphin wearable acoustic interfaces; generative audio augmentation; coral reef acoustic monitoring; echolocation as communication; reference-free translator validation; machine translation with large language models
| title | author | length |
|---|---|---|
| WhAM: Towards A Translative Model of Sperm Whale Vocalization | Orr Paradise | 53m |
A generative pipeline converts arbitrary audio into sperm-whale–style codas using a VampNet-style audio-to-audio model with a Descript VQ‑VAE tokenizer and a… | ||
| Overview and Marine Bioacoustics and Perch Model | Tom Denton, Lauren Harrell | 39m |
A system uses a pre-trained embedding model to embed millions of hours (Australian Acoustic Observatory: 360 microphones, ~3M hours) and then performs search… | ||
| Machine Translation of Human Languages in the Age of LLMs: Is This the End of the Language Barrier? | Markus Freitag, Hadar Shemtov | 57m |
Large LLMs now drive machine translation research and production, yielding near-human scores for many head languages while still failing on the long tail… | ||
| What Makes a Good Language? | David Peterson, Jessie Peterson | 59m |
Overview of concrete methods and results: constructed languages appear as full systems with made proto-forms and applied regular sound changes to generate… | ||
| Decoding Dolphin Speech | Mark Hamilton | 01h 07m |
A system adapts ImageBind-style multimodal models to force alignment of visual spatial tokens with audio by adding a region-matching loss and emergent dual… | ||
| Long term work and Machine Learning Approaches to DolphinGemma | Denise Herzing, Thad E. Starner | 01h 00m |
Four decades of annotated underwater video and high‑frequency audio from Bahamas Atlantic spotted dolphins form the training set and permit correlating… | ||
| Self-Supervised Learning in Transformer Models for Birdsong | Tim Gardener | 53m |
A custom bidirectional transformer trains on spectrograms with no tokenization, uses masked reconstruction as self‑supervision, and produces latent embeddings… | ||
| Restoring Coral Reefs at Scale: AI and the Next Generation of Conservation | Rich Williams | 50m |
Coral Gardeners is running deployable field tech: an iPhone data app called Reef app, a solar-powered, connected reef cam with stereo cameras and hydrophone… | ||
| Validation by Listening | Adam Tauman Kalai | 56m |
Claims a reference-free ET shuffling method can validate translators using only target text, tests it with GPT-4.1/3.5 on real and synthetic conlangs including… | ||
| Potential of Echolocation for Communication | Yossi Yovel, Goldshtein et al. | 54m |
Claims that echolocation doubles as communication rest on concrete recordings and tests: GPS + on-bat microphones track dozens of individuals, a sensory-motor… | ||
| nothing matches | ||