Language models (LMs) have been argued to overlap substantially with human beings in grammaticality judgment tasks. But when humans systematically make errors in language processing, should we expect LMs to behave like cognitive models of language and mimic human behavior? We answer this question by investigating LMs' more subtle judgments associated with "language illusions" -- sentences that are vague in meaning, implausible, or ungrammatical but receive unexpectedly high acceptability judgments by humans. We looked at three illusions: the comparative illusion (e.g. "More people have been to Russia than I have"), the depth-charge illusion (e.g. "No head injury is too trivial to be ignored"), and the negative polarity item (NPI) illusion (e.g. "The hunter who no villager believed to be trustworthy will ever shoot a bear"). We found that probabilities represented by LMs were more likely to align with human judgments of being "tricked" by the NPI illusion which examines a structural dependency, compared to the comparative and the depth-charge illusions which require sophisticated semantic understanding. No single LM or metric yielded results that are entirely consistent with human behavior. Ultimately, we show that LMs are limited both in their construal as cognitive models of human language processing and in their capacity to recognize nuanced but critical information in complicated language materials.

The remarkable capability of large language models (LLMs) in generating high-quality code has drawn increasing attention in the software testing community. However, existing code LLMs often demonstrate unsatisfactory capabilities in generating accurate and complete tests since they were trained on code snippets collected without differentiating between code for testing purposes and other code. In this paper, we present a large-scale dataset UniTSyn, which is capable of enhancing the prowess of LLMs for Unit Test Synthesis. Associating tests with the tested functions is crucial for LLMs to infer the expected behavior and the logic paths to be verified. By leveraging Language Server Protocol, UniTSyn achieves the challenging goal of collecting focal-test pairs without per-project execution setups or per-language heuristics that tend to be fragile and difficult to scale. It contains 2.7 million focal-test pairs across five mainstream programming languages, making it possible to be utilized for enhancing the test generation ability of LLMs. The details of UniTSyn can be found in Table 1. Our experiments demonstrate that, by building an autoregressive model based on UniTSyn, we can achieve significant benefits in learning and understanding unit test representations, resulting in improved generation accuracy and code coverage across all evaluated programming languages. Code and data will be publicly available.

Electroencephalography (EEG) plays a vital role in recording brain activities and is integral to the development of brain-computer interface (BCI) technologies. However, the limited availability and high variability of EEG signals present substantial challenges in creating reliable BCIs. To address this issue, we propose a practical solution drawing on the latest developments in deep learning and Wasserstein Generative Adversarial Network (WGAN). The WGAN was trained on the BCI2000 dataset, consisting of around 1500 EEG recordings and 64 channels from 45 individuals. The generated EEG signals were evaluated via three classifiers yielding improved average accuracies. The quality of generated signals measured using Frechet Inception Distance (FID) yielded scores of 1.345 and 11.565 for eyes-open and closed respectively. Even without a spectral or spatial loss term, our WGAN model was able to emulate the spectral and spatial properties of the EEG training data. The WGAN-generated data mirrored the dominant alpha activity during closed-eye resting and high delta waves in the training data in its topographic map and power spectral density (PSD) plot. Our research testifies to the potential of WGANs in addressing the limited EEG data issue for BCI development by enhancing a small dataset to improve classifier generalizability.

Recent studies confirm the applicability of Inertial Measurement Unit (IMU)-based systems for human motion analysis. Notwithstanding, high-end IMU-based commercial solutions are yet too expensive and complex to democratize their use among a wide range of potential users. Less featured entry-level commercial solutions are being introduced in the market, trying to fill this gap, but still present some limitations that need to be overcome. At the same time, there is a growing number of scientific papers using not commercial, but custom do-it-yourself IMU-based systems in medical and sports applications. Even though these solutions can help to popularize the use of this technology, they have more limited features and the description on how to design and build them from scratch is yet too scarce in the literature. The aim of this work is two-fold: (1) Proving the feasibility of building an affordable custom solution aimed at simultaneous multiple body parts orientation tracking; while providing a detailed bottom-up description of the required hardware, tools, and mathematical operations to estimate and represent 3D movement in real-time. (2) Showing how the introduction of a custom 2.4 GHz communication protocol including a channel hopping strategy can address some of the current communication limitations of entry-level commercial solutions. The proposed system can be used for wireless real-time human body parts orientation tracking with up to 10 custom sensors, at least at 50 Hz. In addition, it provides a more reliable motion data acquisition in Bluetooth and Wi-Fi crowded environments, where the use of entry-level commercial solutions might be unfeasible. This system can be used as a groundwork for developing affordable human motion analysis solutions that do not require an accurate kinematic analysis.

Substantial advances in multi-modal Artificial Intelligence (AI) facilitate the combination of diverse medical modalities to achieve holistic health assessments. We present COMPRER , a novel multi-modal, multi-objective pretraining framework which enhances medical-image representation, diagnostic inferences, and prognosis of diseases. COMPRER employs a multi-objective training framework, where each objective introduces distinct knowledge to the model. This includes a multimodal loss that consolidates information across different imaging modalities; A temporal loss that imparts the ability to discern patterns over time; Medical-measure prediction adds appropriate medical insights; Lastly, reconstruction loss ensures the integrity of image structure within the latent space. Despite the concern that multiple objectives could weaken task performance, our findings show that this combination actually boosts outcomes on certain tasks. Here, we apply this framework to both fundus images and carotid ultrasound, and validate our downstream tasks capabilities by predicting both current and future cardiovascular conditions. COMPRER achieved higher Area Under the Curve (AUC) scores in evaluating medical conditions compared to existing models on held-out data. On the Out-of-distribution (OOD) UK-Biobank dataset COMPRER maintains favorable performance over well-established models with more parameters, even though these models were trained on $75\times$ more data than COMPRER. In addition, to better assess our model's performance in contrastive learning, we introduce a novel evaluation metric, providing deeper understanding of the effectiveness of the latent space pairing.

Education is a dynamic field that must be adaptable to sudden changes and disruptions caused by events like pandemics, war, and natural disasters related to climate change. When these events occur, traditional classrooms with traditional or blended delivery can shift to fully online learning, which requires an efficient learning environment that meets students' needs. While learning management systems support teachers' productivity and creativity, they typically provide the same content to all learners in a course, ignoring their unique learning styles. To address this issue, we propose a semi-supervised machine learning approach that detects students' learning styles using a data mining technique. We use the commonly used Felder Silverman learning style model and demonstrate that our semi-supervised method can produce reliable classification models with few labeled data. We evaluate our approach on two different courses and achieve an accuracy of 88.83% and 77.35%, respectively. Our work shows that educational data mining and semi-supervised machine learning techniques can identify different learning styles and create a personalized learning environment.

In this study, we explore efficient simulation implementations to demonstrate computational equivalence across various models of autonomous mobile robot swarms. Our focus is on Rsynch, a scheduler designed for energy-restricted robots, which falls between Fsynch and Ssynch. We propose efficient protocols for simulating n(>=2) luminous (LUMI) robots operating in Rsynch using LUMI robots in Ssynch or Asynch. Our contributions are twofold: (1) We introduce protocols that simulate LUMI robots in Rsynch using 4k colors in Ssynch and 5k colors in Asynch, for algorithms that employ k colors. This approach notably reduces the number of colors needed for Ssynch simulations of Rsynch, compared to previous efforts. Meanwhile, the color requirement for Asynch simulations remains consistent with previous Asynch simulations of Ssynch, facilitating the simulation of Rsynch in Asynch. (2) We establish that for n=2, Rsynch can be optimally simulated in Asynch using a minimal number of colors. Additionally, we confirm that all our proposed simulation protocols are self-stabilizing, ensuring functionality from any initial configuration.

In an era where artificial intelligence (AI) intertwines with medical research, the delineation of truth becomes increasingly complex. This study ostensibly examines a purported novel SARS-CoV-2 variant, dubbed the Omega variant, showcasing 31 unique mutations in the S gene region. However, the real undercurrent of this narrative is a demonstration of the ease with which AI, specifically ChatGPT-4, can fabricate convincing yet entirely fictional scientific data. The so-called Omega variant was identified in a fully vaccinated, previously infected 35-year-old male presenting with severe COVID-19 symptoms. Through a detailed, albeit artificial, genomic analysis and contact tracing, this study mirrors the rigorous methodology of genuine case reports, thereby setting the stage for a compelling but entirely constructed narrative. The entire case study was generated by ChatGPT-4, a large language model by OpenAI. The fabricated Omega variant features an ensemble of mutations, including N501Y and E484K, known for enhancing ACE2 receptor affinity, alongside L452R and P681H, ostensibly indicative of immune evasion. This variant's contrived interaction dynamics - severe symptoms in a vaccinated individual versus mild ones in unvaccinated contacts - were designed to mimic real-world complexities, including suggestions of antibody-dependent enhancement (ADE). While the Omega variant is a product of AI-generated fiction, the implications of this exercise are real and profound. The ease with which AI can generate believable but false scientific information, as illustrated in this case, raises significant concerns about the potential for misinformation in medicine. This study, therefore, serves as a cautionary tale, emphasizing the necessity for critical evaluation of sources, especially in an age where AI tools like ChatGPT are becoming increasingly sophisticated and widespread in their use.

One-on-one tutoring is an effective instructional method for enhancing learning, yet its efficacy hinges on tutor competencies. Novice math tutors often prioritize content-specific guidance, neglecting aspects such as social-emotional learning. Social-emotional learning promotes equity and inclusion and nurturing relationships with students, which is crucial for holistic student development. Assessing the competencies of tutors accurately and efficiently can drive the development of tailored tutor training programs. However, evaluating novice tutor ability during real-time tutoring remains challenging as it typically requires experts-in-the-loop. To address this challenge, this preliminary study aims to harness Generative Pre-trained Transformers (GPT), such as GPT-3.5 and GPT-4 models, to automatically assess tutors' ability of using social-emotional tutoring strategies. Moreover, this study also reports on the financial dimensions and considerations of employing these models in real-time and at scale for automated assessment. The current study examined four prompting strategies: two basic Zero-shot prompt strategies, Tree of Thought prompt, and Retrieval-Augmented Generator (RAG) based prompt. The results indicate that the RAG prompt demonstrated more accurate performance (assessed by the level of hallucination and correctness in the generated assessment texts) and lower financial costs than the other strategies evaluated. These findings inform the development of personalized tutor training interventions to enhance the the educational effectiveness of tutored learning.

Compared with gradient based artificial neural networks, biological neural networks usually show a more powerful generalization ability to quickly adapt to unknown environments without using any gradient back-propagation procedure. Inspired by the distributed memory mechanism of human brains, we propose a novel gradient-free Distributed Memorization Learning mechanism, namely DML, to support quick domain adaptation of transferred models. In particular, DML adopts randomly connected neurons to memorize the association of input signals, which are propagated as impulses, and makes the final decision by associating the distributed memories based on their confidence. More importantly, DML is able to perform reinforced memorization based on unlabeled data to quickly adapt to a new domain without heavy fine-tuning of deep features, which makes it very suitable for deploying on edge devices. Experiments based on four cross-domain real-world datasets show that DML can achieve superior performance of real-time domain adaptation compared with traditional gradient based MLP with more than 10% improvement of accuracy while reducing 87% of the timing cost of optimization.