Video Experimental Relacionado
Updated: Jan 8, 2026

03:14
Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness
Published on: December 6, 2024
986
No Hay Almuerzo Gratis: La Generación Aumentada por Recuperación Socava la Equidad en los LLM, Incluso para Usuarios
Mengxuan Hu1, Hongyi Wu1, Ronghang Zhu2
1University of Virginia.
Resumen
La Generación Aumentada por Recuperación (RAG) puede introducir problemas de equidad en los modelos de lenguaje grandes (LLM) incluso con datos censurados. Esta investigación revela que RAG
Área de la Ciencia:
- Inteligencia Artificial
- Procesamiento del Lenguaje Natural
- Ética del Aprendizaje Automático
Sus antecedentes:
- La Generación Aumentada por Recuperación (RAG) es una técnica popular para mejorar los Modelos de Lenguaje Grandes (LLM).
- RAG mejora la facticidad y el conocimiento específico del dominio de los LLM sin un reentrenamiento costoso.
- Persisten las preocupaciones sobre la equidad y los posibles sesgos en los resultados de los LLM.
Objetivo del estudio:
- Investigar los costos y las implicaciones de la equidad del uso de RAG.
- Analizar cómo los diferentes niveles de censura de datos en RAG afectan los resultados de equidad.
- Evaluar la vulnerabilidad de los sistemas RAG a ataques relacionados con la equidad.
Principales métodos:
- Se propuso un modelo de amenaza de tres niveles basado en la conciencia de la equidad del usuario.
- Se realizaron experimentos utilizando RAG con conjuntos de datos externos sin censura, parcialmente censurados y completamente censurados.
- Se evaluaron las implicaciones de equidad sin ajustar ni reentrenar los LLM.
Principales resultados:
- La alineación de la equidad en RAG se puede comprometer fácilmente sin modificación del modelo.
- Incluso los conjuntos de datos externos completamente censurados y supuestamente imparciales pueden generar resultados sesgados de LLM a través de RAG.
- Los sistemas RAG demuestran susceptibilidad a la socavación de la equidad.
Conclusiones:
- Los métodos actuales de alineación de LLM pueden ser insuficientes para los sistemas RAG.
- RAG introduce desafíos de equidad únicos que requieren estrategias de mitigación novedosas.
- Es crucial una mayor investigación para desarrollar salvaguardias de equidad sólidas en los LLM basados en RAG.
Videos de Conceptos Relacionados
Retrieval
387
Retrieval is the process of getting information out of memory storage and back into conscious awareness. This ability is essential for daily tasks like brushing hair and teeth, driving to work, and performing job duties. Retrieval occurs in three ways: recall, recognition, and relearning.
Recall involves accessing information without cues, such as during an essay test, where individuals must retrieve facts and concepts from memory unaided. Another example is remembering the name of a colleague...
Recall involves accessing information without cues, such as during an essay test, where individuals must retrieve facts and concepts from memory unaided. Another example is remembering the name of a colleague...
387
Survival Tree
369
Survival trees are a non-parametric method used in survival analysis to model the relationship between a set of covariates and the time until an event of interest occurs, often referred to as the "time-to-event" or "survival time." This method is particularly useful when dealing with censored data, where the event has not occurred for some individuals by the end of the study period, or when the exact time of the event is unknown.
Building a Survival Tree
Constructing a...
Building a Survival Tree
Constructing a...
369
Improving Translational Accuracy
14.0K
Base complementarity between the three base pairs of mRNA codon and the tRNA anticodon is not a failsafe mechanism. Inaccuracies can range from a single mismatch to no correct base pairing at all. The free energy difference between the correct and nearly correct base pairs can be as small as 3 kcal/ mol. With complementarity being the only proofreading step, the estimated error frequency would be one wrong amino acid in every 100 amino acids incorporated. However, error frequencies observed in...
14.0K
Improving Translational Accuracy
3.5K
3.5K
Language and Cognition
688
Language serves as a bridge between ideas and communication, influencing how individuals perceive and interact with the world. Psychologists have long debated whether language shapes thought or vice versa. This discussion gained grip with Edward Sapir and Benjamin Lee Whorf in the 1940s, who proposed that language determines thought, a concept known as linguistic determinism. They suggested that the vocabulary and structure of a language influence how its speakers think and perceive reality.
688
Language Development
804
Children master language quickly and with relative ease, supported by both biological predisposition and reinforcement. B. F. Skinner (1957) proposed that language is learned through reinforcement, while Noam Chomsky (1965) argued that language acquisition mechanisms are biologically determined.
The critical period for language acquisition suggests that the ability to acquire language is at its peak early in life. As people age, this proficiency decreases. Language development begins very...
The critical period for language acquisition suggests that the ability to acquire language is at its peak early in life. As people age, this proficiency decreases. Language development begins very...
804
