AI opens new era in cognitive studies of wild primates

6 hours ago

Scientists developed CapuchinAI, an AI system that uses facial recognition and real-time touchscreen testing to automate cognitive studies of wild capuchin monkeys.
The system identifies individual capuchins with 97% accuracy and presents learning tasks tailored to each monkey, rewarding correct responses with food.
Field tests in Costa Rica showed rapid habituation and learning, demonstrating a scalable method for cognitive testing in natural environments.
The open-source paper provides a blueprint for building a low-cost, field-ready platform, aiming to bridge the gap between lab experiments and wild observations.
The system captures individual differences in cognitive abilities, allowing researchers to link life history data to cognition and explore how environment shapes behavior.
CapuchinAI is designed to be adaptable to other wild primate species with recorded life histories, expanding its potential for comparative cognition studies.

Researchers demonstrated that quantum computers can enhance the accuracy and scope of generative AI models for drug discovery.
A hybrid quantum-classical system accelerated AI to generate novel peptides, improving binding prediction to target proteins, especially with scarce training data.
The project was self-funded by the team using spare time and leftover money, as innovative science is often deemed risky by traditional funders.

Conventional AI continuously processes data even when nothing changes, leading to high energy waste.
Researchers developed a cerebellum-inspired memtransistor that ignores expected inputs and quickly detects unexpected events with significantly lower energy use.
The device identified abnormal heart rhythms within one-fifth of a heartbeat with over 98% accuracy, using about 10,000 times fewer operations than conventional AI.

Schema harness enables frontier models to achieve ~99% performance on ARC-AGI-3 Public by treating latent world representations as executable programs, allowing interpretable, verifiable, and searchable modeling.
ARC-AGI-3 challenges agents to infer game rules, objects, and goals from raw 64x64 color grids without explicit instructions, using the Relative Human Action Efficiency (RHAE) metric for evaluation.
The approach combines state grounding (identifying objects and variables from observations) and mechanism discovery (learning transition rules), solved jointly within an editable program that mirrors physicist-like reasoning.

Brain2Qwerty v2 is a non-invasive AI system that decodes brain activity into text in real-time, achieving high accuracy without surgical implants.
The system was trained on 22,000 sentences from nine participants using MEG recordings and end-to-end deep learning, leveraging large language models for semantic context.
It achieved a 61% word accuracy rate overall and 78% for the best participant, significantly outperforming other non-invasive methods (8% accuracy).

Cultural battles over AI have intensified due to Big AI's lack of consent, environmental impact, and closed systems, while critics are dismissed by zealots making extreme claims.
A small cohort explores what 'good' AI would look like: consensual data, open source, low hardware requirements, and creator-controlled, exemplified by Corridor Digital's CorridorKey tool.
CorridorKey is an open-source green screen keying tool trained on CGI-generated data, built with help from commercial AI, and released for free with community improvements.

Researchers demonstrated that quantum computers can enhance the accuracy and scope of generative AI models for drug discovery.
A hybrid quantum-classical system accelerated AI to generate novel peptides, improving binding prediction to target proteins, especially with scarce training data.
The project was self-funded by the team using spare time and leftover money, as innovative science is often deemed risky by traditional funders.

Conventional AI continuously processes data even when nothing changes, leading to high energy waste.
Researchers developed a cerebellum-inspired memtransistor that ignores expected inputs and quickly detects unexpected events with significantly lower energy use.
The device identified abnormal heart rhythms within one-fifth of a heartbeat with over 98% accuracy, using about 10,000 times fewer operations than conventional AI.

Schema harness enables frontier models to achieve ~99% performance on ARC-AGI-3 Public by treating latent world representations as executable programs, allowing interpretable, verifiable, and searchable modeling.
ARC-AGI-3 challenges agents to infer game rules, objects, and goals from raw 64x64 color grids without explicit instructions, using the Relative Human Action Efficiency (RHAE) metric for evaluation.
The approach combines state grounding (identifying objects and variables from observations) and mechanism discovery (learning transition rules), solved jointly within an editable program that mirrors physicist-like reasoning.

Brain2Qwerty v2 is a non-invasive AI system that decodes brain activity into text in real-time, achieving high accuracy without surgical implants.
The system was trained on 22,000 sentences from nine participants using MEG recordings and end-to-end deep learning, leveraging large language models for semantic context.
It achieved a 61% word accuracy rate overall and 78% for the best participant, significantly outperforming other non-invasive methods (8% accuracy).

Cultural battles over AI have intensified due to Big AI's lack of consent, environmental impact, and closed systems, while critics are dismissed by zealots making extreme claims.
A small cohort explores what 'good' AI would look like: consensual data, open source, low hardware requirements, and creator-controlled, exemplified by Corridor Digital's CorridorKey tool.
CorridorKey is an open-source green screen keying tool trained on CGI-generated data, built with help from commercial AI, and released for free with community improvements.

Hasty Briefsbeta