Large Language Models (LLMs) combined with instruction tuning have made significant progress when generalizing to unseen tasks. However, they have been less successful in Information Extraction (IE), lagging behind task-specific models. Typically, IE tasks are characterized by complex annotation guidelines which describe the task and give examples to humans. Previous attempts to leverage such information have failed, even with the largest models, as they are not able to follow the guidelines out-of-the-box. In this paper we propose GoLLIE (Guideline-following Large Language Model for IE), a model able to improve zero-shot results on unseen IE tasks by virtue of being fine-tuned to comply with annotation guidelines. Comprehensive evaluation empirically demonstrates that GoLLIE is able to generalize to and follow unseen guidelines, outperforming previous attempts at zero-shot information extraction. The ablation study shows that detailed guidelines is key for good results.
相關內容
Recent advancements have significantly augmented the reasoning capabilities of Large Language Models (LLMs) through various methodologies, especially chain-of-thought (CoT) reasoning. However, previous methods fail to address reasoning errors in intermediate steps, leading to accumulative errors.In this paper, we propose Deductive Beam Search (DBS), which seamlessly integrates CoT and deductive reasoning with step-wise beam search for LLMs. Our approach deploys a verifier, verifying the deducibility of a reasoning step and its premises, thus alleviating the error accumulation. Furthermore, we introduce a scalable and labor-free data construction method to amplify our model's verification capabilities. Extensive experiments demonstrate that our approach significantly enhances the base performance of LLMs of various scales (7B, 13B, 70B, and ChatGPT) across 8 reasoning datasets from 3 diverse reasoning genres, including arithmetic, commonsense, and symbolic. Moreover, our analysis proves DBS's capability of detecting diverse and subtle reasoning errors and robustness on different model scales.
While there has been significant development of models for Plain Language Summarization (PLS), evaluation remains a challenge. PLS lacks a dedicated assessment metric, and the suitability of text generation evaluation metrics is unclear due to the unique transformations involved (e.g., adding background explanations, removing specialized terminology). To address these concerns, our study presents a granular meta-evaluation testbed, APPLS, designed to evaluate metrics for PLS. We define a set of perturbations along four criteria inspired by previous work that a PLS metric should capture: informativeness, simplification, coherence, and faithfulness. An analysis of metrics using our testbed reveals that current metrics fail to capture simplification consistently. In response, we introduce POMME, a new metric designed to assess text simplification in PLS; the metric is calculated as the normalized perplexity difference between an in-domain and out-of-domain language model. We demonstrate POMME's correlation with fine-grained variations in simplification and validate its sensitivity across 4 text simplification datasets. This work contributes the first meta-evaluation testbed for PLS and a comprehensive evaluation of existing metrics. The APPLS testbed and POMME is available at //github.com/LinguisticAnomalies/APPLS.
The field of Neural Style Transfer (NST) has witnessed remarkable progress in the past few years, with approaches being able to synthesize artistic and photorealistic images and videos of exceptional quality. To evaluate such results, a diverse landscape of evaluation methods and metrics is used, including authors' opinions based on side-by-side comparisons, human evaluation studies that quantify the subjective judgements of participants, and a multitude of quantitative computational metrics which objectively assess the different aspects of an algorithm's performance. However, there is no consensus regarding the most suitable and effective evaluation procedure that can guarantee the reliability of the results. In this review, we provide an in-depth analysis of existing evaluation techniques, identify the inconsistencies and limitations of current evaluation methods, and give recommendations for standardized evaluation practices. We believe that the development of a robust evaluation framework will not only enable more meaningful and fairer comparisons among NST methods but will also enhance the comprehension and interpretation of research findings in the field.
Multi-modal Large Language Models (MLLMs) have shown impressive abilities in generating reasonable responses with respect to multi-modal contents. However, there is still a wide gap between the performance of recent MLLM-based applications and the expectation of the broad public, even though the most powerful OpenAI's GPT-4 and Google's Gemini have been deployed. This paper strives to enhance understanding of the gap through the lens of a qualitative study on the generalizability, trustworthiness, and causal reasoning capabilities of recent proprietary and open-source MLLMs across four modalities: ie, text, code, image, and video, ultimately aiming to improve the transparency of MLLMs. We believe these properties are several representative factors that define the reliability of MLLMs, in supporting various downstream applications. To be specific, we evaluate the closed-source GPT-4 and Gemini and 6 open-source LLMs and MLLMs. Overall we evaluate 230 manually designed cases, where the qualitative results are then summarized into 12 scores (ie, 4 modalities times 3 properties). In total, we uncover 14 empirical findings that are useful to understand the capabilities and limitations of both proprietary and open-source MLLMs, towards more reliable downstream multi-modal applications.
Large pretrained language models are widely used in downstream NLP tasks via task-specific fine-tuning, but such procedures can be costly. Recently, Parameter-Efficient Fine-Tuning (PEFT) methods have achieved strong task performance while updating much fewer parameters than full model fine-tuning (FFT). However, it is non-trivial to make informed design choices on the PEFT configurations, such as their architecture, the number of tunable parameters, and even the layers in which the PEFT modules are inserted. Consequently, it is highly likely that the current, manually designed configurations are suboptimal in terms of their performance-efficiency trade-off. Inspired by advances in neural architecture search, we propose AutoPEFT for automatic PEFT configuration selection: we first design an expressive configuration search space with multiple representative PEFT modules as building blocks. Using multi-objective Bayesian optimisation in a low-cost setup, we then discover a Pareto-optimal set of configurations with strong performance-cost trade-offs across different numbers of parameters that are also highly transferable across different tasks. Empirically, on GLUE and SuperGLUE tasks, we show that AutoPEFT-discovered configurations significantly outperform existing PEFT methods and are on par or better than FFT without incurring substantial training efficiency costs.
Spin Transfer Torque Random Access Memory (STT-RAM) is an emerging Non-Volatile Memory (NVM) technology that has garnered attention to overcome the drawbacks of conventional CMOS-based technologies. However, such technologies must be evaluated before deployment under real workloads and architecture. But there is a lack of available open-source STT-RAM-based system evaluation framework, which hampers research and experimentation and impacts the adoption of STT- RAM in a system. This paper proposes a novel, extendable STT-RAM memory controller design integrated inside the gem5 simulator. Our framework enables understanding various aspects of STT-RAM, i.e., power, delay, clock cycles, energy, and system throughput. We will open-source our HOPE framework, which will fuel research and aid in accelerating the development of future system architectures based on STT-RAM. It will also facilitate the user for further tool enhancement.
Recently, Graph Neural Network (GNN)-based vulnerability detection systems have achieved remarkable success. However, the lack of explainability poses a critical challenge to deploy black-box models in security-related domains. For this reason, several approaches have been proposed to explain the decision logic of the detection model by providing a set of crucial statements positively contributing to its predictions. Unfortunately, due to the weakly-robust detection models and suboptimal explanation strategy, they have the danger of revealing spurious correlations and redundancy issue. In this paper, we propose Coca, a general framework aiming to 1) enhance the robustness of existing GNN-based vulnerability detection models to avoid spurious explanations; and 2) provide both concise and effective explanations to reason about the detected vulnerabilities. \sysname consists of two core parts referred to as Trainer and Explainer. The former aims to train a detection model which is robust to random perturbation based on combinatorial contrastive learning, while the latter builds an explainer to derive crucial code statements that are most decisive to the detected vulnerability via dual-view causal inference as explanations. We apply Coca over three typical GNN-based vulnerability detectors. Experimental results show that Coca can effectively mitigate the spurious correlation issue, and provide more useful high-quality explanations.
Dysarthric speech reconstruction (DSR) systems aim to automatically convert dysarthric speech into normal-sounding speech. The technology eases communication with speakers affected by the neuromotor disorder and enhances their social inclusion. NED-based (Neural Encoder-Decoder) systems have significantly improved the intelligibility of the reconstructed speech as compared with GAN-based (Generative Adversarial Network) approaches, but the approach is still limited by training inefficiency caused by the cascaded pipeline and auxiliary tasks of the content encoder, which may in turn affect the quality of reconstruction. Inspired by self-supervised speech representation learning and discrete speech units, we propose a Unit-DSR system, which harnesses the powerful domain-adaptation capacity of HuBERT for training efficiency improvement and utilizes speech units to constrain the dysarthric content restoration in a discrete linguistic space. Compared with NED approaches, the Unit-DSR system only consists of a speech unit normalizer and a Unit HiFi-GAN vocoder, which is considerably simpler without cascaded sub-modules or auxiliary tasks. Results on the UASpeech corpus indicate that Unit-DSR outperforms competitive baselines in terms of content restoration, reaching a 28.2% relative average word error rate reduction when compared to original dysarthric speech, and shows robustness against speed perturbation and noise.
We introduce the concept of "Alternative Speech" as a new way to directly combat hate speech and complement the limitations of counter-narrative. An alternative speech provides practical alternatives to hate speech in real-world scenarios by offering speech-level corrections to speakers while considering the surrounding context and promoting speakers to reform. Further, an alternative speech can combat hate speech alongside counter-narratives, offering a useful tool to address social issues such as racial discrimination and gender inequality. We propose the new concept and provide detailed guidelines for constructing the necessary dataset. Through discussion, we demonstrate that combining alternative speech and counter-narrative can be a more effective strategy for combating hate speech by complementing specificity and guiding capacity of counter-narrative. This paper presents another perspective for dealing with hate speech, offering viable remedies to complement the constraints of current approaches to mitigating harmful bias.
Understanding the decision-making process of Graph Neural Networks (GNNs) is crucial to their interpretability. Most existing methods for explaining GNNs typically rely on training auxiliary models, resulting in the explanations remain black-boxed. This paper introduces Graph Output Attribution (GOAt), a novel method to attribute graph outputs to input graph features, creating GNN explanations that are faithful, discriminative, as well as stable across similar samples. By expanding the GNN as a sum of scalar products involving node features, edge features and activation patterns, we propose an efficient analytical method to compute contribution of each node or edge feature to each scalar product and aggregate the contributions from all scalar products in the expansion form to derive the importance of each node and edge. Through extensive experiments on synthetic and real-world data, we show that our method not only outperforms various state-ofthe-art GNN explainers in terms of the commonly used fidelity metric, but also exhibits stronger discriminability, and stability by a remarkable margin.
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