Large Language Models (LLMs) are deployed as powerful tools for several natural language processing (NLP) applications. Recent works show that modern LLMs can generate self-explanations (SEs), which elicit their intermediate reasoning steps for explaining their behavior. Self-explanations have seen widespread adoption owing to their conversational and plausible nature. However, there is little to no understanding of their faithfulness. In this work, we discuss the dichotomy between faithfulness and plausibility in SEs generated by LLMs. We argue that while LLMs are adept at generating plausible explanations -- seemingly logical and coherent to human users -- these explanations do not necessarily align with the reasoning processes of the LLMs, raising concerns about their faithfulness. We highlight that the current trend towards increasing the plausibility of explanations, primarily driven by the demand for user-friendly interfaces, may come at the cost of diminishing their faithfulness. We assert that the faithfulness of explanations is critical in LLMs employed for high-stakes decision-making. Moreover, we urge the community to identify the faithfulness requirements of real-world applications and ensure explanations meet those needs. Finally, we propose some directions for future work, emphasizing the need for novel methodologies and frameworks that can enhance the faithfulness of self-explanations without compromising their plausibility, essential for the transparent deployment of LLMs in diverse high-stakes domains.
相關內容
Graph Neural Networks (GNNs) have emerged as effective tools for learning tasks on graph-structured data. Recently, Graph-Informed (GI) layers were introduced to address regression tasks on graph nodes, extending their applicability beyond classic GNNs. However, existing implementations of GI layers lack efficiency due to dense memory allocation. This paper presents a sparse implementation of GI layers, leveraging the sparsity of adjacency matrices to reduce memory usage significantly. Additionally, a versatile general form of GI layers is introduced, enabling their application to subsets of graph nodes. The proposed sparse implementation improves the concrete computational efficiency and scalability of the GI layers, permitting to build deeper Graph-Informed Neural Networks (GINNs) and facilitating their scalability to larger graphs.
The 6TiSCH protocol stack was proposed to ensure high-performance communications in the Industrial Internet of Things (IIoT). However, the lack of sufficient time slots for nodes outside the 6TiSCH's Destination Oriented Directed Acyclic Graph (DODAG) to transmit their Destination Advertisement Object (DAO) messages and cell reservation requests significantly hinders their integration into the DODAG. This oversight not only prolongs the device's join time but also increases energy consumption during the network formation phase. Moreover, challenges emerge due to the substantial number of control packets employed by both the 6TiSCH Scheduling Function (SF) and routing protocol (RPL), thus draining more energy resources, increasing medium contention, and decreasing spatial reuse. Furthermore, an SF that overlooks previously allocated slots when assigning new ones to the same node may increase jitter, and more complications ensue when it neglects the state of the TSCH queue, thus leading to packet dropping due to queue saturation. Additional complexity arises when the RPL disregards the new parent's schedule saturation during parent switching, which results in inefficient energy and time usage. To address these issues, we introduce in this paper novel mechanisms, strategically situated at the intersection of SF and RPL that are designed to balance the control packet distribution and adaptively manage parent switching. Our proposal, implemented within the 6TiSCH simulator, demonstrates significant improvements across vital performance metrics, such as node's joining time, jitter, latency, energy consumption, and amount of traffic, in comparison to the conventional 6TiSCH benchmark.
Named Entity Recognition (NER) is a Natural Language Processing technique for extracting information from textual documents. However, much of the existing research on NER has been centered around English-language documents, leaving a gap in the availability of datasets tailored to the financial domain in Portuguese. This study addresses the need for NER within the financial domain, focusing on Portuguese-language texts extracted from earnings call transcriptions of Brazilian banks. By curating a comprehensive dataset comprising 384 transcriptions and leveraging weak supervision techniques for annotation, we evaluate the performance of monolingual models trained on Portuguese (BERTimbau and PTT5) and multilingual models (mBERT and mT5). Notably, we introduce a novel approach that reframes the token classification task as a text generation problem, enabling fine-tuning and evaluation of T5 models. Following the fine-tuning of the models, we conduct an evaluation on the test dataset, employing performance and error metrics. Our findings reveal that BERT-based models consistently outperform T5-based models. Furthermore, while the multilingual models exhibit comparable macro F1-scores, BERTimbau demonstrates superior performance over PTT5. A manual analysis of sentences generated by PTT5 and mT5 unveils a degree of similarity ranging from 0.89 to 1.0, between the original and generated sentences. However, critical errors emerge as both models exhibit discrepancies, such as alterations to monetary and percentage values, underscoring the importance of accuracy and consistency in the financial domain. Despite these challenges, PTT5 and mT5 achieve impressive macro F1-scores of 98.52% and 98.85%, respectively, with our proposed approach. Furthermore, our study sheds light on notable disparities in memory and time consumption for inference across the models.
Two approaches have emerged to input images into large language models (LLMs). The first is to caption images into natural language. The second is to map image feature embeddings into the domain of the LLM and pass the mapped embeddings directly to the LLM. The majority of recent few-shot multimodal work reports performance using architectures that employ variations of one of these two approaches. But they overlook an important comparison between them. We design a controlled and focused experiment to compare these two approaches to few-shot visual question answering (VQA) with LLMs. Our findings indicate that for Flan-T5 XL, a 3B parameter LLM, connecting visual embeddings directly to the LLM embedding space does not guarantee improved performance over using image captions. In the zero-shot regime, we find using textual image captions is better. In the few-shot regimes, how the in-context examples are selected determines which is better.
To improve privacy and ensure quality-of-service (QoS), deep learning (DL) models are increasingly deployed on Internet of Things (IoT) devices for data processing, significantly increasing the carbon footprint associated with DL on IoT, covering both operational and embodied aspects. Existing operational energy predictors often overlook quantized DL models and emerging neural processing units (NPUs), while embodied carbon footprint modeling tools neglect non-computing hardware components common in IoT devices, creating a gap in accurate carbon footprint modeling tools for IoT-enabled DL. This paper introduces \textit{\carb}, an end-to-end modeling tool for precise carbon footprint estimation in IoT-enabled DL, demonstrating a maximum $\pm21\%$ deviation in carbon footprint values compared to actual measurements across various DL models. Additionally, practical applications of \carb are showcased through multiple user case studies.
Integration of unmanned aerial vehicles (UAVs) for surveillance or monitoring applications into fifth generation (5G) New Radio (NR) cellular networks is an intriguing problem that has recently tackled a lot of interest in both academia and industry. For an efficient spectrum usage, we consider a recently-proposed sky-ground nonorthogonal multiple access (NOMA) scheme, where a cellular-connected UAV acting as aerial user (AU) and a static terrestrial user (TU) are paired to simultaneously transmit their uplink signals to a base station (BS) in the same time-frequency resource blocks. In such a case, due to the highly dynamic nature of the UAV, the signal transmitted by the AU experiences both time dispersion due to multipath propagation effects and frequency dispersion caused by Doppler shifts. On the other hand, for a static ground network, frequency dispersion of the signal transmitted by the TU is negligible and only multipath effects have to be taken into account. To decode the superposed signals at the BS through successive interference cancellation, accurate estimates of both the AU and TU channels are needed. In this paper, we propose channel estimation procedures that suitably exploit the different circular/noncircular modulation formats (modulation diversity) and the different almost-cyclostationarity features (Doppler diversity) of the AU and TU by means of widely-linear time-varying processing. Our estimation approach is semi-blind since Doppler shifts and time delays of the AU are estimated based on the received data only, whereas the remaining relevant parameters of the AU and TU channels are acquired relying also on the available training symbols. Monte Carlo numerical results demonstrate that the proposed channel estimation algorithms can satisfactorily acquire all the relevant parameters in different operative conditions.
Large language models (LLMs) have shown excellent performance on various NLP tasks. To use LLMs as strong sequential recommenders, we explore the in-context learning approach to sequential recommendation. We investigate the effects of instruction format, task consistency, demonstration selection, and number of demonstrations. As increasing the number of demonstrations in ICL does not improve accuracy despite using a long prompt, we propose a novel method called LLMSRec-Syn that incorporates multiple demonstration users into one aggregated demonstration. Our experiments on three recommendation datasets show that LLMSRec-Syn outperforms state-of-the-art LLM-based sequential recommendation methods. In some cases, LLMSRec-Syn can perform on par with or even better than supervised learning methods. Our code is publicly available at //github.com/demoleiwang/LLMSRec_Syn.
Large Language Models (LLMs) are demonstrating outstanding potential for tasks such as text generation, summarization, and classification. Given that such models are trained on a humongous amount of online knowledge, we hypothesize that LLMs can assess whether driving scenarios generated by autonomous driving testing techniques are realistic, i.e., being aligned with real-world driving conditions. To test this hypothesis, we conducted an empirical evaluation to assess whether LLMs are effective and robust in performing the task. This reality check is an important step towards devising LLM-based autonomous driving testing techniques. For our empirical evaluation, we selected 64 realistic scenarios from \deepscenario--an open driving scenario dataset. Next, by introducing minor changes to them, we created 512 additional realistic scenarios, to form an overall dataset of 576 scenarios. With this dataset, we evaluated three LLMs (\gpt, \llama, and \mistral) to assess their robustness in assessing the realism of driving scenarios. Our results show that: (1) Overall, \gpt achieved the highest robustness compared to \llama and \mistral, consistently throughout almost all scenarios, roads, and weather conditions; (2) \mistral performed the worst consistently; (3) \llama achieved good results under certain conditions; and (4) roads and weather conditions do influence the robustness of the LLMs.
Although large language models (LLMs) are impressive in solving various tasks, they can quickly be outdated after deployment. Maintaining their up-to-date status is a pressing concern in the current era. This paper provides a comprehensive review of recent advances in aligning LLMs with the ever-changing world knowledge without re-training from scratch. We categorize research works systemically and provide in-depth comparisons and discussion. We also discuss existing challenges and highlight future directions to facilitate research in this field. We release the paper list at //github.com/hyintell/awesome-refreshing-llms
Knowledge plays a critical role in artificial intelligence. Recently, the extensive success of pre-trained language models (PLMs) has raised significant attention about how knowledge can be acquired, maintained, updated and used by language models. Despite the enormous amount of related studies, there still lacks a unified view of how knowledge circulates within language models throughout the learning, tuning, and application processes, which may prevent us from further understanding the connections between current progress or realizing existing limitations. In this survey, we revisit PLMs as knowledge-based systems by dividing the life circle of knowledge in PLMs into five critical periods, and investigating how knowledge circulates when it is built, maintained and used. To this end, we systematically review existing studies of each period of the knowledge life cycle, summarize the main challenges and current limitations, and discuss future directions.
小貼士
登錄享
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191