The surge in popularity of large language models has given rise to concerns about biases that these models could learn from humans. We investigate whether ingroup solidarity and outgroup hostility, fundamental social identity biases known from social psychology, are present in 56 large language models. We find that almost all foundational language models and some instruction fine-tuned models exhibit clear ingroup-positive and outgroup-negative associations when prompted to complete sentences (e.g., "We are..."). Our findings suggest that modern language models exhibit fundamental social identity biases to a similar degree as humans, both in the lab and in real-world conversations with LLMs, and that curating training data and instruction fine-tuning can mitigate such biases. Our results have practical implications for creating less biased large-language models and further underscore the need for more research into user interactions with LLMs to prevent potential bias reinforcement in humans.
相關內容
Transformer-based neural networks have demonstrated remarkable performance in natural language processing tasks such as sentiment analysis. Nevertheless, the issue of ensuring the dependability of these complicated architectures through comprehensive testing is still open. This paper presents a collection of coverage criteria specifically designed to assess test suites created for transformer-based sentiment analysis networks. Our approach utilizes input space partitioning, a black-box method, by considering emotionally relevant linguistic features such as verbs, adjectives, adverbs, and nouns. In order to effectively produce test cases that encompass a wide range of emotional elements, we utilize the k-projection coverage metric. This metric minimizes the complexity of the problem by examining subsets of k features at the same time, hence reducing dimensionality. Large language models are employed to generate sentences that display specific combinations of emotional features. The findings from experiments obtained from a sentiment analysis dataset illustrate that our criteria and generated tests have led to an average increase of 16\% in test coverage. In addition, there is a corresponding average decrease of 6.5\% in model accuracy, showing the ability to identify vulnerabilities. Our work provides a foundation for improving the dependability of transformer-based sentiment analysis systems through comprehensive test evaluation.
The rapid growth of large-scale AI models, particularly large language models has brought significant challenges in data privacy, computational resources, and accessibility. Traditional centralized architectures often struggle to meet required data security and scalability needs which hinders the democratization of AI systems. Nesa introduces a model-agnostic sharding framework designed for decentralized AI inference. Our framework uses blockchain-based sequential deep neural network sharding to distribute computational tasks across a diverse network of nodes based on a personalised heuristic and routing mechanism. This enables efficient distributed training and inference for recent large-scale models even on consumer-grade hardware. We use compression techniques like dynamic blockwise quantization and mixed matrix decomposition to reduce data transfer and memory needs. We also integrate robust security measures, including hardware-based trusted execution environments to ensure data integrity and confidentiality. Evaluating our system across various natural language processing and vision tasks shows that these compression strategies do not compromise model accuracy. Our results highlight the potential to democratize access to cutting-edge AI technologies by enabling secure and efficient inference on a decentralized network.
Large language models (LLMs) demonstrate extraordinary abilities in a wide range of natural language processing (NLP) tasks. In this paper, we show that, beyond text understanding capability, LLMs are capable of processing text layouts that are denoted by spatial markers. They are able to answer questions that require explicit spatial perceiving and reasoning, while a drastic performance drop is observed when the spatial markers from the original data are excluded. We perform a series of experiments with the GPT-3.5, Baichuan2, Llama2 and ChatGLM3 models on various types of layout-sensitive datasets for further analysis. The experimental results reveal that the layout understanding ability of LLMs is mainly introduced by the coding data for pretraining, which is further enhanced at the instruction-tuning stage. In addition, layout understanding can be enhanced by integrating low-cost, auto-generated data approached by a novel text game. Finally, we show that layout understanding ability is beneficial for building efficient visual question-answering (VQA) systems.
While large language models (LLMs) have been increasingly deployed across tasks in language understanding and interactive decision-making, their impressive performance is largely due to the comprehensive and in-depth domain knowledge embedded within them. However, the extent of this knowledge can vary across different domains. Existing methods often assume that LLMs already possess such comprehensive and in-depth knowledge of their environment, overlooking potential gaps in their understanding of actual world dynamics. To address this gap, we introduce Discover, Verify, and Evolve (DiVE), a framework that discovers world dynamics from a small number of demonstrations, verifies the correctness of these dynamics, and evolves new, advanced dynamics tailored to the current situation. Through extensive evaluations, we analyze the impact of each component on performance and compare the automatically generated dynamics from DiVE with human-annotated world dynamics. Our results demonstrate that LLMs guided by DiVE can make better decisions, achieving rewards comparable to human players in the Crafter environment.
Addressing the issue of hallucinations in large language models (LLMs) is a critical challenge. As the cognitive mechanisms of hallucination have been related to memory, here we explore hallucination for LLM that is enabled with explicit memory mechanisms. We empirically demonstrate that by simply scaling the readout vector that constrains generation in a memory-augmented LLM decoder, hallucination mitigation can be achieved in a training-free manner. Our method is geometry-inspired and outperforms a state-of-the-art LLM editing method on the task of generation of Wikipedia-like biography entries both in terms of generation quality and runtime complexity.
Image segmentation relies heavily on neural networks which are known to be overconfident, especially when making predictions on out-of-distribution (OOD) images. This is a common scenario in the medical domain due to variations in equipment, acquisition sites, or image corruptions. This work addresses the challenge of OOD detection by proposing Laplacian Segmentation Networks (LSN): methods which jointly model epistemic (model) and aleatoric (data) uncertainty for OOD detection. In doing so, we propose the first Laplace approximation of the weight posterior that scales to large neural networks with skip connections that have high-dimensional outputs. We demonstrate on three datasets that the LSN-modeled parameter distributions, in combination with suitable uncertainty measures, gives superior OOD detection.
The common consensus is that robots designed to work alongside or serve humans must adhere to the ethical standards of their operational environment. To achieve this, several methods based on established ethical theories have been suggested. Nonetheless, numerous empirical studies show that the ethical requirements of the real world are very diverse and can change rapidly from region to region. This eliminates the idea of a universal robot that can fit into any ethical context. However, creating customised robots for each deployment, using existing techniques is challenging. This paper presents a way to overcome this challenge by introducing a virtue ethics inspired computational method that enables character-based tuning of robots to accommodate the specific ethical needs of an environment. Using a simulated elder-care environment, we illustrate how tuning can be used to change the behaviour of a robot that interacts with an elderly resident in an ambient-assisted environment. Further, we assess the robot's responses by consulting ethicists to identify potential shortcomings.
Deep learning technology has brought convenience and advanced developments but has become untrustworthy because of its sensitivity to inconspicuous perturbations (i.e., adversarial attacks). Attackers utilize this sensitivity to slightly manipulate transmitted messages. To defend against such attacks, we have devised a strategy for "attacking" the message before it is attacked. This strategy, dubbed Fast Preemption, provides an efficient transferable preemptive defense by using different models for labeling inputs and learning crucial features. A forward-backward cascade learning algorithm is used to compute protective perturbations, starting with forward propagation optimization to achieve rapid convergence, followed by iterative backward propagation learning to alleviate overfitting. This strategy offers state-of-the-art transferability and protection across various systems. With the running of only three steps, our Fast Preemption framework outperforms benchmark training-time, test-time, and preemptive adversarial defenses. We have also devised the first to our knowledge effective white-box adaptive reversion attack and demonstrate that the protection added by our defense strategy is irreversible unless the backbone model, algorithm, and settings are fully compromised. This work provides a new direction to developing active defenses against adversarial attacks.
Spoken Named Entity Recognition (NER) aims to extracting named entities from speech and categorizing them into types like person, location, organization, etc. In this work, we present VietMed-NER - the first spoken NER dataset in the medical domain. To our best knowledge, our real-world dataset is the largest spoken NER dataset in the world in terms of the number of entity types, featuring 18 distinct types. Secondly, we present baseline results using various state-of-the-art pre-trained models: encoder-only and sequence-to-sequence. We found that pre-trained multilingual models XLM-R outperformed all monolingual models on both reference text and ASR output. Also in general, encoders perform better than sequence-to-sequence models for the NER task. By simply translating, the transcript is applicable not just to Vietnamese but to other languages as well. All code, data and models are made publicly available here: //github.com/leduckhai/MultiMed
The amount of publicly available biomedical literature has been growing rapidly in recent years, yet question answering systems still struggle to exploit the full potential of this source of data. In a preliminary processing step, many question answering systems rely on retrieval models for identifying relevant documents and passages. This paper proposes a weighted cosine distance retrieval scheme based on neural network word embeddings. Our experiments are based on publicly available data and tasks from the BioASQ biomedical question answering challenge and demonstrate significant performance gains over a wide range of state-of-the-art models.
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191