Llama
LLaMa
Simulating Thought Disorder: Fine-Tuning Llama-2 for Synthetic Speech in Schizophrenia
Llama 3.1 Paper: The Llama Family of Models
Modern artificial intelligence (AI) systems are powered by foundation models. This paper presents a new set of foundation models, called Llama 3. It is a herd of language models that natively support multilinguality, coding, reasoning, and tool usage. Our largest model is a dense Transformer with 405B parameters and a context window of up to 128K tokens. This paper presents an extensive empirical evaluation of Llama 3. We find that Llama 3 delivers comparable quality to leading language models such as GPT-4 on a plethora of tasks. We publicly release Llama 3, including pre-trained and post-trained versions of the 405B parameter language model and our Llama Guard 3 model for input and output safety. The paper also presents the results of experiments in which we integrate image, video, and speech capabilities into Llama 3 via a compositional approach. We observe this approach performs competitively with the state-of-the-art on image, video, and speech recognition tasks. The resulting models are not yet being broadly released as they are still under development.
Enhancing Qualitative Coding with Large Language Models: Potential and Challenges
Qualitative coding is the process of categorizing and labeling raw data to identify themes, patterns, and concepts within qualitative research. This process requires significant time, reflection, and discussion, often characterized by inherent subjectivity and uncertainty. Here, we explore the possibility to leverage large language models (LLM) to enhance the process and assist researchers with qualitative coding. LLMs, trained on extensive human-generated text, possess an architecture that renders them capable of understanding the broader context of a conversation or text. This allows them to extract patterns and meaning effectively, making them particularly useful for the accurate extraction and coding of relevant themes. In our current approach, we employed the chatGPT 3.5 Turbo API, integrating it into the qualitative coding process for data from the SWISS100 study, specifically focusing on data derived from centenarians' experiences during the Covid-19 pandemic, as well as a systematic centenarian literature review. We provide several instances illustrating how our approach can assist researchers with extracting and coding relevant themes. With data from human coders on hand, we highlight points of convergence and divergence between AI and human thematic coding in the context of these data. Moving forward, our goal is to enhance the prototype and integrate it within an LLM designed for local storage and operation (LLaMa). Our initial findings highlight the potential of AI-enhanced qualitative coding, yet they also pinpoint areas requiring attention. Based on these observations, we formulate tentative recommendations for the optimal integration of LLMs in qualitative coding research. Further evaluations using varied datasets and comparisons among different LLMs will shed more light on the question of whether and how to integrate these models into this domain.
Tools for Analyzing and Repairing the Brain. (Simultaneous translation to Spanish)
To enable the understanding and repair of complex biological systems, such as the brain, we are creating novel optical tools that enable molecular-resolution maps of such systems, as well as technologies for observing and controlling high-speed physiological dynamics in such systems. First, we have developed a method for imaging specimens with nanoscale precision, by embedding them in a swellable polymer, homogenizing their mechanical properties, and exposing them to water – which causes them to expand manyfold isotropically. This method, which we call expansion microscopy (ExM), enables ordinary microscopes to do nanoscale imaging, in a multiplexed fashion – important, for example, for brain mapping. Second, we have developed a set of genetically-encoded reagents, known as optogenetic tools, that when expressed in specific neurons, enable their electrical activities to be precisely driven or silenced in response to millisecond timescale pulses of light. Finally, we are designing, and evolving, novel reagents, such as fluorescent voltage indicators and somatically targeted calcium indicators, to enable the imaging of fast physiological processes in 3-D with millisecond precision. In this way we aim to enable the systematic mapping, control, and dynamical observation of complex biological systems like the brain. The talk will be simultaneously interpreted English-Spanish) by the Interpreter, Mg. Lourdes Martino. Para permitir la comprensión y reparación de sistemas biológicos complejos, como el cerebro, estamos creando herramientas ópticas novedosas que permiten crear mapas de resolución molecular de dichos sistemas, así como tecnologías para observar y controlar la dinámica fisiológica de alta velocidad en dichos sistemas. Primero, hemos desarrollado un método para obtener imágenes de muestras con precisión a nanoescala, incrustándolas en un polímero hinchable, homogeneizando sus propiedades mecánicas y exponiéndolas al agua, lo que hace que se expandan muchas veces isotrópicamente. Este método, que llamamos microscopía de expansión (ExM), permite que los microscopios ordinarios obtengan imágenes a nanoescala, de forma multiplexada, lo que es importante, por ejemplo, para el mapeo cerebral. En segundo lugar, hemos desarrollado un conjunto de reactivos codificados genéticamente, conocidos como herramientas optogenéticas, que cuando se expresan en neuronas específicas, permiten que sus actividades eléctricas sean activadas o silenciadas con precisión en respuesta a pulsos de luz en una escala de tiempo de milisegundos. Finalmente, estamos diseñando y desarrollando reactivos novedosos, como indicadores de voltaje fluorescentes e indicadores de calcio dirigidos somáticamente, para permitir la obtención de imágenes de procesos fisiológicos rápidos en 3-D con precisión de milisegundos. De esta manera, nuestro objetivo es permitir el mapeo sistemático, el control y la observación dinámica de sistemas biológicos complejos como el cerebro. La conferencia será traducida simultáneamente al español por la intérprete Mg. Lourdes Martino.