Large language models (LLMs) are gaining increasing popularity in both academia and industry, owing to their unprecedented performance in various applications. As LLMs continue to play a vital role in both research and daily use, their evaluation becomes increasingly critical, not only at the task level, but also at the society level for better understanding of their potential risks. Over the past years, significant efforts have been made to examine LLMs from various perspectives. This paper presents a comprehensive review of these evaluation methods for LLMs, focusing on three key dimensions: what to evaluate, where to evaluate, and how to evaluate. Firstly, we provide an overview from the perspective of evaluation tasks, encompassing general natural language processing tasks, reasoning, medical usage, ethics, educations, natural and social sciences, agent applications, and other areas. Secondly, we answer the `where' and `how' questions by diving into the evaluation methods and benchmarks, which serve as crucial components in assessing performance of LLMs. Then, we summarize the success and failure cases of LLMs in different tasks. Finally, we shed light on several future challenges that lie ahead in LLMs evaluation. Our aim is to offer invaluable insights to researchers in the realm of LLMs evaluation, thereby aiding the development of more proficient LLMs. Our key point is that evaluation should be treated as an essential discipline to better assist the development of LLMs. We consistently maintain the related open-source materials at: //github.com/MLGroupJLU/LLM-eval-survey.
相關內容
Satellite Image Time Series (SITS) representation learning is complex due to high spatiotemporal resolutions, irregular acquisition times, and intricate spatiotemporal interactions. These challenges result in specialized neural network architectures tailored for SITS analysis. The field has witnessed promising results achieved by pioneering researchers, but transferring the latest advances or established paradigms from Computer Vision (CV) to SITS is still highly challenging due to the existing suboptimal representation learning framework. In this paper, we develop a novel perspective of SITS processing as a direct set prediction problem, inspired by the recent trend in adopting query-based transformer decoders to streamline the object detection or image segmentation pipeline. We further propose to decompose the representation learning process of SITS into three explicit steps: collect-update-distribute, which is computationally efficient and suits for irregularly-sampled and asynchronous temporal satellite observations. Facilitated by the unique reformulation, our proposed temporal learning backbone of SITS, initially pre-trained on the resource efficient pixel-set format and then fine-tuned on the downstream dense prediction tasks, has attained new state-of-the-art (SOTA) results on the PASTIS benchmark dataset. Specifically, the clear separation between temporal and spatial components in the semantic/panoptic segmentation pipeline of SITS makes us leverage the latest advances in CV, such as the universal image segmentation architecture, resulting in a noticeable 2.5 points increase in mIoU and 8.8 points increase in PQ, respectively, compared to the best scores reported so far.
While Multimodal Large Language Models (MLLMs) are widely used for a variety of vision-language tasks, one observation is that they sometimes misinterpret visual inputs or fail to follow textual instructions even in straightforward cases, leading to irrelevant responses, mistakes, and ungrounded claims. This observation is analogous to a phenomenon in neuropsychology known as Agnosia, an inability to correctly process sensory modalities and recognize things (e.g., objects, colors, relations). In our study, we adapt this similar concept to define "agnosia in MLLMs", and our goal is to comprehensively evaluate and mitigate such agnosia in MLLMs. Inspired by the diagnosis and treatment process in neuropsychology, we propose a novel framework EMMA (Evaluation and Mitigation of Multimodal Agnosia). In EMMA, we develop an evaluation module that automatically creates fine-grained and diverse visual question answering examples to assess the extent of agnosia in MLLMs comprehensively. We also develop a mitigation module to reduce agnosia in MLLMs through multimodal instruction tuning on fine-grained conversations. To verify the effectiveness of our framework, we evaluate and analyze agnosia in seven state-of-the-art MLLMs using 9K test samples. The results reveal that most of them exhibit agnosia across various aspects and degrees. We further develop a fine-grained instruction set and tune MLLMs to mitigate agnosia, which led to notable improvement in accuracy.
Large language models, particularly those akin to the rapidly progressing GPT series, are gaining traction for their expansive influence. While there is keen interest in their applicability within medical domains such as psychology, tangible explorations on real-world data remain scant. Concurrently, users on social media platforms are increasingly vocalizing personal sentiments; under specific thematic umbrellas, these sentiments often manifest as negative emotions, sometimes escalating to suicidal inclinations. Timely discernment of such cognitive distortions and suicidal risks is crucial to effectively intervene and potentially avert dire circumstances. Our study ventured into this realm by experimenting on two pivotal tasks: suicidal risk and cognitive distortion identification on Chinese social media platforms. Using supervised learning as a baseline, we examined and contrasted the efficacy of large language models via three distinct strategies: zero-shot, few-shot, and fine-tuning. Our findings revealed a discernible performance gap between the large language models and traditional supervised learning approaches, primarily attributed to the models' inability to fully grasp subtle categories. Notably, while GPT-4 outperforms its counterparts in multiple scenarios, GPT-3.5 shows significant enhancement in suicide risk classification after fine-tuning. To our knowledge, this investigation stands as the maiden attempt at gauging large language models on Chinese social media tasks. This study underscores the forward-looking and transformative implications of using large language models in the field of psychology. It lays the groundwork for future applications in psychological research and practice.
Quantitative Transmission Electron Microscopy (TEM) during in-situ straining experiment is able to reveal the motion of dislocations -- linear defects in the crystal lattice of metals. In the domain of materials science, the knowledge about the location and movement of dislocations is important for creating novel materials with superior properties. A long-standing problem, however, is to identify the position and extract the shape of dislocations, which would ultimately help to create a digital twin of such materials. In this work, we quantitatively compare state-of-the-art instance segmentation methods, including Mask R-CNN and YOLOv8. The dislocation masks as the results of the instance segmentation are converted to mathematical lines, enabling quantitative analysis of dislocation length and geometry -- important information for the domain scientist, which we then propose to include as a novel length-aware quality metric for estimating the network performance. Our segmentation pipeline shows a high accuracy suitable for all domain-specific, further post-processing. Additionally, our physics-based metric turns out to perform much more consistently than typically used pixel-wise metrics.
Large language models (LLMs), such as ChatGPT and GPT-4, are versatile and can solve different tasks due to their emergent ability and generalizability. However, LLMs sometimes lack domain-specific knowledge to perform tasks, which would also cause hallucination during inference. In some previous works, additional modules like graph neural networks (GNNs) are trained on retrieved knowledge from external knowledge bases, aiming to mitigate the problem of lacking domain-specific knowledge. However, incorporating additional modules: 1) would need retraining additional modules when encountering novel domains; 2) would become a bottleneck since LLMs' strong abilities are not fully utilized for retrieval. In this paper, we propose a paradigm, termed Knowledge Solver (KSL), to teach LLMs to search for essential knowledge from external knowledge bases by harnessing their own strong generalizability. Specifically, we design a simple yet effective prompt to transform retrieval into a multi-hop decision sequence, which empowers LLMs with searching knowledge ability in zero-shot manner. Additionally, KSL is able to provide complete retrieval paths and therefore increase explainability of LLMs' reasoning processes. We conduct experiments on three datasets: CommonsenseQA, OpenbookQA, and MedQA-USMLE, and found that our approach improves LLM baseline performance by a relatively large margin.
Deep neural networks (DNNs) have made great strides in pushing the state-of-the-art in several challenging domains. Recent studies reveal that they are prone to making overconfident predictions. This greatly reduces the overall trust in model predictions, especially in safety-critical applications. Early work in improving model calibration employs post-processing techniques which rely on limited parameters and require a hold-out set. Some recent train-time calibration methods, which involve all model parameters, can outperform the postprocessing methods. To this end, we propose a new train-time calibration method, which features a simple, plug-and-play auxiliary loss known as multi-class alignment of predictive mean confidence and predictive certainty (MACC). It is based on the observation that a model miscalibration is directly related to its predictive certainty, so a higher gap between the mean confidence and certainty amounts to a poor calibration both for in-distribution and out-of-distribution predictions. Armed with this insight, our proposed loss explicitly encourages a confident (or underconfident) model to also provide a low (or high) spread in the presoftmax distribution. Extensive experiments on ten challenging datasets, covering in-domain, out-domain, non-visual recognition and medical image classification scenarios, show that our method achieves state-of-the-art calibration performance for both in-domain and out-domain predictions. Our code and models will be publicly released.
Graph neural networks (GNNs) have been demonstrated to be a powerful algorithmic model in broad application fields for their effectiveness in learning over graphs. To scale GNN training up for large-scale and ever-growing graphs, the most promising solution is distributed training which distributes the workload of training across multiple computing nodes. However, the workflows, computational patterns, communication patterns, and optimization techniques of distributed GNN training remain preliminarily understood. In this paper, we provide a comprehensive survey of distributed GNN training by investigating various optimization techniques used in distributed GNN training. First, distributed GNN training is classified into several categories according to their workflows. In addition, their computational patterns and communication patterns, as well as the optimization techniques proposed by recent work are introduced. Second, the software frameworks and hardware platforms of distributed GNN training are also introduced for a deeper understanding. Third, distributed GNN training is compared with distributed training of deep neural networks, emphasizing the uniqueness of distributed GNN training. Finally, interesting issues and opportunities in this field are discussed.
Knowledge graphs represent factual knowledge about the world as relationships between concepts and are critical for intelligent decision making in enterprise applications. New knowledge is inferred from the existing facts in the knowledge graphs by encoding the concepts and relations into low-dimensional feature vector representations. The most effective representations for this task, called Knowledge Graph Embeddings (KGE), are learned through neural network architectures. Due to their impressive predictive performance, they are increasingly used in high-impact domains like healthcare, finance and education. However, are the black-box KGE models adversarially robust for use in domains with high stakes? This thesis argues that state-of-the-art KGE models are vulnerable to data poisoning attacks, that is, their predictive performance can be degraded by systematically crafted perturbations to the training knowledge graph. To support this argument, two novel data poisoning attacks are proposed that craft input deletions or additions at training time to subvert the learned model's performance at inference time. These adversarial attacks target the task of predicting the missing facts in knowledge graphs using KGE models, and the evaluation shows that the simpler attacks are competitive with or outperform the computationally expensive ones. The thesis contributions not only highlight and provide an opportunity to fix the security vulnerabilities of KGE models, but also help to understand the black-box predictive behaviour of KGE models.
In pace with developments in the research field of artificial intelligence, knowledge graphs (KGs) have attracted a surge of interest from both academia and industry. As a representation of semantic relations between entities, KGs have proven to be particularly relevant for natural language processing (NLP), experiencing a rapid spread and wide adoption within recent years. Given the increasing amount of research work in this area, several KG-related approaches have been surveyed in the NLP research community. However, a comprehensive study that categorizes established topics and reviews the maturity of individual research streams remains absent to this day. Contributing to closing this gap, we systematically analyzed 507 papers from the literature on KGs in NLP. Our survey encompasses a multifaceted review of tasks, research types, and contributions. As a result, we present a structured overview of the research landscape, provide a taxonomy of tasks, summarize our findings, and highlight directions for future work.
Recent developments in image classification and natural language processing, coupled with the rapid growth in social media usage, have enabled fundamental advances in detecting breaking events around the world in real-time. Emergency response is one such area that stands to gain from these advances. By processing billions of texts and images a minute, events can be automatically detected to enable emergency response workers to better assess rapidly evolving situations and deploy resources accordingly. To date, most event detection techniques in this area have focused on image-only or text-only approaches, limiting detection performance and impacting the quality of information delivered to crisis response teams. In this paper, we present a new multimodal fusion method that leverages both images and texts as input. In particular, we introduce a cross-attention module that can filter uninformative and misleading components from weak modalities on a sample by sample basis. In addition, we employ a multimodal graph-based approach to stochastically transition between embeddings of different multimodal pairs during training to better regularize the learning process as well as dealing with limited training data by constructing new matched pairs from different samples. We show that our method outperforms the unimodal approaches and strong multimodal baselines by a large margin on three crisis-related tasks.
小貼士
登錄享
相關主題
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191