Field Notes: Defining Decode

The workshop underscored that we are on the precipice of something big: more research focus, data, and technology are being directed towards animal communication than ever before. As we start to uncover more about our planet and its inhabitants, our understanding of other species will fundamentally change, enabling humans to reimagine our relationships with the natural world.

Read More

Introducing BirdAVES: Self-Supervised Audio Foundation Model for Birds

BirdAVES is a series of newly trained animal vocalization encoder models developed using self-supervision. It represents an over 20% improvement in bird-related datasets and tasks.

Read More

Using AI to Understand Animal Communication

New funding from Paul G Allen Family Foundation to support development of technologies that will reconnect us to nature 

Human beings are part of nature. But the vast majority of what is happening in this complex, interconnected natural world we inhabit is often beyond our ability to perceive - let alone comprehend. This disconnection from the world around us arguably represents one of the biggest obstacles humanity faces in tackling the biodiversity and climate crises. 

Fortunately the exponential progress we’re seeing in AI offers new ways of looking at the world -  extending our ability to perceive through technology, and finding patterns in data that will help to make sense of what is happening around us.

At Earth Species Project we are focused on harnessing these new developments in AI to gain a better understanding of animal communication because we believe it offers a unique and valuable way to connect to the rest of nature. 

We’re thrilled to announce that we have just received significant new funding of $1.2 million from the Paul G Allen Family Foundation that will support these efforts.

Read More

Voxaboxen: New Tool Supports Annotation of Large Audio Files

ESP Senior Research Scientists Benjamin Hoffman and Maddie Cusimano announce the release of Voxaboxen, a new machine learning tool for research in bioacoustics. Based upon Earth Species Project's bioacoustics foundation model AVES, Voxaboxen is designed to detect and classify animal vocalizations in recorded audio, with the high level of temporal specificity that is required for studies focused on communication behavior. In this blog post, we will give an overview of why and how we designed Voxaboxen, as well as how to apply Voxaboxen to your own data. We anticipate this new tool will support and advance research by providing a straightforward way to automate annotation of large audio files.

Read More

New grants to catalyze research into AI and non-human communication

While scientists have been studying the behavior and communication of other species for decades, the use of AI to help accelerate these discoveries is relatively new. In partnership with FootPrint Coalition (FPC), Robert Downey Jr’s initiative to accelerate technologies for sustainable abundance, and the Experiment Foundation, ESP is thrilled to announce the launch of a new grant program designed to catalyze new and emerging research into interspecies communication. 

Read More

Finding the signal in the noise: How machine learning can help us perceive, understand, and protect other species

Sara Keen, Senior Researcher, Behavioral Ecology and AI reflects on how machine learning can help us find the signals in the noise and decipher meaning from the world of invisible information that surrounds us.

Read More

ESP Technical Roadmap

This blog provides an overview of ESP's technical roadmap - offering a path forward for our proposed and ongoing scientific work, which is focused on using AI to decode on non-human animal communication.  

Machine learning (ML) has proven to be a powerful tool for learning latent representations of data (LeCun et al., 2015). For example, learned representations of human languages have enabled translation between different languages without the use of dictionaries (Artetxe et al., 2017; Lample et al., 2017). More recently, ML has been able to generate realistic images based on text descriptions (DALL-E 2), as well as predict and imitate the words we speak (AudioLM).

We hypothesize that ML models can provide new insights into non-human communication, and believe that discoveries fueled by these new techniques have the potential to transform our relationship with the rest of nature. We also expect that many of the techniques we develop will be useful tools for processing data in applied conservation settings (Tuia et al., 2022).

Read More

Using AI to Listen to All of Earth's Species: New insights from the field

Whispering whales, turtle hatchlings communicating with each other through their shells, coral larvae that can hear and gravitate toward their home reefs…

Director of Development and External Affairs Jane Lawton reflects on the insights shared by Dr. Karen Bakker, Dr. Ari Friedlaender and Aza Raskin at ESP's October 25 panel discussion on interspecies communication in San Francisco - from how sophisticated the communication systems of other species are, to the explosion of new bioacoustic technologies and AI that see us on the cusp of dramatically expanding our understanding of those communication systems. And simultaneously exposing how little we, as human beings, currently know.

Read More

Turning Motion into Meaning: the ESP journey to becoming Nat Geo Explorers

Katie Zacarian, CEO, Earth Species Project, reflects on the ESP journey to receiving a National Geographic Explorer grant. The project we have received funding for is focused on developing machine learning models to interpret animal motion data gathered on powerful animal-borne sensors, and it represents an important step on the journey toward decoding animal communication.

Read More

Self-Supervised Ethogram Discovery

Senior AI research scientist Benjamin Hoffman is working on self-supervised methods for interpreting data collected from animal-borne tags, known as bio-loggers. Using bio-loggers, scientists are able to record an animal’s motions, as well as audio and video footage from the animal’s perspective. However, these data are often difficult to interpret, and there is often too much data to analyze by hand. A solution is to use self-supervised learning to discover repeated behavioral patterns in these data.

Read More