Recently, high-performing code generation systems based on large language models have surfaced. They are trained on massive corpora containing much more natural text than actual executable computer code. This work shows that current code generation systems exhibit undesired biases inherited from their large language model backbones, which can reduce the quality of the generated code under specific circumstances. To investigate the effect, we propose the "block of influence" concept, which enables a modular decomposition and analysis of the coding challenges. We introduce an automated intervention mechanism reminiscent of adversarial testing that exposes undesired biases through the failure modes of the models under test. Finally, we demonstrate how our framework can be used as a data transformation technique during fine-tuning, acting as a mitigation strategy for these biases.

Span-based models are one of the most straightforward methods for named entity recognition (NER). Existing span-based NER systems shallowly aggregate the token representations to span representations. However, this typically results in significant ineffectiveness for long-span entities, a coupling between the representations of overlapping spans, and ultimately a performance degradation. In this study, we propose DSpERT (Deep Span Encoder Representations from Transformers), which comprises a standard Transformer and a span Transformer. The latter uses low-layered span representations as queries, and aggregates the token representations as keys and values, layer by layer from bottom to top. Thus, DSpERT produces span representations of deep semantics. With weight initialization from pretrained language models, DSpERT achieves performance higher than or competitive with recent state-of-the-art systems on eight NER benchmarks. Experimental results verify the importance of the depth for span representations, and show that DSpERT performs particularly well on long-span entities and nested structures. Further, the deep span representations are well structured and easily separable in the feature space.

Task-oriented communications, mostly using learning-based joint source-channel coding (JSCC), aim to design a communication-efficient edge inference system by transmitting task-relevant information to the receiver. However, only transmitting task-relevant information without introducing any redundancy may cause robustness issues in learning due to the channel variations, and the JSCC which directly maps the source data into continuous channel input symbols poses compatibility issues on existing digital communication systems. In this paper, we address these two issues by first investigating the inherent tradeoff between the informativeness of the encoded representations and the robustness to information distortion in the received representations, and then propose a task-oriented communication scheme with digital modulation, named discrete task-oriented JSCC (DT-JSCC), where the transmitter encodes the features into a discrete representation and transmits it to the receiver with the digital modulation scheme. In the DT-JSCC scheme, we develop a robust encoding framework, named robust information bottleneck (RIB), to improve the communication robustness to the channel variations, and derive a tractable variational upper bound of the RIB objective function using the variational approximation to overcome the computational intractability of mutual information. The experimental results demonstrate that the proposed DT-JSCC achieves better inference performance than the baseline methods with low communication latency, and exhibits robustness to channel variations due to the applied RIB framework.

Recent advancements in Large Language Models (LLMs) harness linguistic associations in vast natural language data for practical applications. However, their ability to understand the physical world using only language data remains a question. After reviewing existing protocols, we explore this question using a novel and tightly controlled reasoning test (ART) and compare human norms against versions of GPT-3. Our findings highlight the categories of common-sense relations models that could learn directly from data and areas of weakness. GPT-3 offers evidence for verbal reasoning on a par with human subjects for several relations including Synonymy, Antonymy, and Default inheritance, Without reinforcement learning from human judgements, it appears GPT-3 performs at the lower end of the reference interval for Has-part and Contained-in. Weaknesses were observed also in affordance characteristics through Necessary-quality, Order-of-size and Order-of-intensity. Combining LLMs with symbolic world grounding is a promising direction to address associative learning.

As robots have become increasingly common in human-rich environments, it is critical that they are able to exhibit social cues to be perceived as a cooperative and socially-conformant team member. We investigate the effect of robot gaze cues on people's subjective perceptions of a mobile robot as a socially present entity in three common hallway navigation scenarios. The tested robot gaze behaviors were path-oriented (looking at its own future path), or person-oriented (looking at the nearest person), with fixed-gaze as the control. We conduct a real-world study with 36 participants who walked through the hallway, and an online study with 233 participants who were shown simulated videos of the same scenarios. Our results suggest that the preferred gaze behavior is scenario-dependent. Person-oriented gaze behaviors which acknowledge the presence of the human are generally preferred when the robot and human cross paths. However, this benefit is diminished in scenarios that involve less implicit interaction between the robot and the human.

Position embeddings, encoding the positional relationships among tokens in text sequences, make great contributions to modeling local context features in Transformer-based pre-trained language models. However, in Extractive Question Answering, position embeddings trained with instances of varied context lengths may not perform well as we expect. Since the embeddings of rear positions are updated fewer times than the front position embeddings, the rear ones may not be properly trained. In this paper, we propose a simple but effective strategy, Random Padding, without any modifications to architectures of existing pre-trained language models. We adjust the token order of input sequences when fine-tuning, to balance the number of updating times of every position embedding. Experiments show that Random Padding can significantly improve model performance on the instances whose answers are located at rear positions, especially when models are trained on short contexts but evaluated on long contexts. Our code and data will be released for future research.

Relation extraction (RE) is the core NLP task of inferring semantic relationships between entities from text. Standard supervised RE techniques entail training modules to tag tokens comprising entity spans and then predict the relationship between them. Recent work has instead treated the problem as a \emph{sequence-to-sequence} task, linearizing relations between entities as target strings to be generated conditioned on the input. Here we push the limits of this approach, using larger language models (GPT-3 and Flan-T5 large) than considered in prior work and evaluating their performance on standard RE tasks under varying levels of supervision. We address issues inherent to evaluating generative approaches to RE by doing human evaluations, in lieu of relying on exact matching. Under this refined evaluation, we find that: (1) Few-shot prompting with GPT-3 achieves near SOTA performance, i.e., roughly equivalent to existing fully supervised models; (2) Flan-T5 is not as capable in the few-shot setting, but supervising and fine-tuning it with Chain-of-Thought (CoT) style explanations (generated via GPT-3) yields SOTA results. We release this model as a new baseline for RE tasks.

The conversational artificial-intelligence (AI) technology ChatGPT has become one of the most widely used natural language processing tools. With thousands of published papers demonstrating its applications across various industries and fields, ChatGPT has sparked significant interest in the research community. Reviews of primary data have also begun to emerge. An overview of the available evidence from multiple reviews and studies could provide further insights, minimize redundancy, and identify areas where further research is needed. Objective: To evaluate the existing reviews and literature related to ChatGPT's applications and its potential impact on different fields by conducting a systematic review of reviews and bibliometric analysis of primary literature. Methods: PubMed, EuropePMC, Dimensions AI, medRxiv, bioRxiv, arXiv, and Google Scholar were searched for ChatGPT-related publications from 2022 to 4/30/2023. Studies including secondary data related to the application of ChatGPT were considered. Reporting and risk of bias assesment was performed using PRISMA guidelines. Results: A total of 305 unique records with potential relevance to the review were identified from a pool of over 2,000 original articles. After multi-step screening process, 11 reviews were selected, consisting of 9 reviews specifically focused on ChatGPT and 2 reviews on broader AI topics that also included discussions on ChatGPT. We also conducted bibliometric analysis of primary data. Conclusions: While AI has the potential to revolutionize various industries, further interdisciplinary research, customized integrations, and ethical innovation are necessary to address existing concerns and ensure its responsible use. Protocol Registration: PROSPERO registration no. CRD42023417336, DOI 10.17605/OSF.IO/87U6Q.

Beyond the success story of adversarial training (AT) in the recent text domain on top of pre-trained language models (PLMs), our empirical study showcases the inconsistent gains from AT on some tasks, e.g. commonsense reasoning, named entity recognition. This paper investigates AT from the perspective of the contextualized language representation outputted by PLM encoders. We find the current AT attacks lean to generate sub-optimal adversarial examples that can fool the decoder part but have a minor effect on the encoder. However, we find it necessary to effectively deviate the latter one to allow AT to gain. Based on the observation, we propose simple yet effective \textit{Contextualized representation-Adversarial Training} (CreAT), in which the attack is explicitly optimized to deviate the contextualized representation of the encoder. It allows a global optimization of adversarial examples that can fool the entire model. We also find CreAT gives rise to a better direction to optimize the adversarial examples, to let them less sensitive to hyperparameters. Compared to AT, CreAT produces consistent performance gains on a wider range of tasks and is proven to be more effective for language pre-training where only the encoder part is kept for downstream tasks. We achieve the new state-of-the-art performances on a series of challenging benchmarks, e.g. AdvGLUE (59.1 $ \rightarrow $ 61.1), HellaSWAG (93.0 $ \rightarrow $ 94.9), ANLI (68.1 $ \rightarrow $ 69.3).

The demand for intelligent industries and smart services based on big data is rising rapidly with the increasing digitization and intelligence of the modern world. This survey comprehensively reviews Blockchained Federated Learning (BlockFL) that joins the benefits of both Blockchain and Federated Learning to provide a secure and efficient solution for the demand. We compare the existing BlockFL models in four Internet-of-Things (IoT) application scenarios: Personal IoT (PIoT), Industrial IoT (IIoT), Internet of Vehicles (IoV), and Internet of Health Things (IoHT), with a focus on security and privacy, trust and reliability, efficiency, and data heterogeneity. Our analysis shows that the features of decentralization and transparency make BlockFL a secure and effective solution for distributed model training, while the overhead and compatibility still need further study. It also reveals the unique challenges of each domain presents unique challenges, e.g., the requirement of accommodating dynamic environments in IoV and the high demands of identity and permission management in IoHT, in addition to some common challenges identified, such as privacy, resource constraints, and data heterogeneity. Furthermore, we examine the existing technologies that can benefit BlockFL, thereby helping researchers and practitioners to make informed decisions about the selection and development of BlockFL for various IoT application scenarios.