Search

The aim of this study is to examine the correlation between reading culture and the mathematical literacy self-efficacy among pre-service teachers and assesses these variables in relation to several factors. Utilizing convenience sampling, the research sample consists of 562 pre-service teachers studying in mathematics and primary school teaching programs in seven different universities. Data was gathered using the Reading Culture and Mathematical Literacy Self-Efficacy Scales, along with a personal information form. Both descriptive and inferential statistical methods were employed to analyze the data. Results revealed a significant but low-level correlation between reading culture, its sub-dimensions, and mathematical literacy self-efficacy levels among pre-service teachers. Notably, pre-service teachers engaging with scientific books demonstrated higher mathematical literacy self-efficacy than those preferring literary books. In addition, pre-service teachers' scores of mathematical literacy self-efficacy and reading culture are not independent of the department they study in. Reading culture, gender, and department were significant predictors of mathematical literacy. [This paper was published in the "International Journal of Progressive Education" v20 n4 p1-30 2024.]

The advancement of Large Language Models (LLMs) and their finetuning strategies has triggered the renewed interests in multi-agent reinforcement learning. In this paper, we introduce a focal diversity-optimized multi-agent reinforcement learning approach, coined as MARL-Focal, with three unique characteristics. First, we develop an agent-fusion framework for encouraging multiple LLM based agents to collaborate in producing the final inference output for each LLM query. Second, we develop a focal-diversity optimized agent selection algorithm that can choose a small subset of the available agents based on how well they can complement one another to generate the query output. Finally, we design a conflict-resolution method to detect output inconsistency among multiple agents and produce our MARL-Focal output through reward-aware and policy-adaptive inference fusion. Extensive evaluations on five benchmarks show that MARL-Focal is cost-efficient and adversarial-robust. Our multi-agent fusion model achieves performance improvement of 5.51\% compared to the best individual LLM-agent and offers stronger robustness over the TruthfulQA benchmark. Code is available at https://github.com/sftekin/rl-focal

Recent research shows that Large Language Models (LLMs) are vulnerable to harmful fine-tuning attacks -- models lose their safety alignment ability after fine-tuning on a few harmful samples. For risk mitigation, a guardrail is typically used to filter out harmful samples before fine-tuning. By designing a new red-teaming method, we in this paper show that purely relying on the moderation guardrail for data filtration is not reliable. Our proposed attack method, dubbed Virus, easily bypasses the guardrail moderation by slightly modifying the harmful data. Experimental results show that the harmful data optimized by Virus is not detectable by the guardrail with up to 100\% leakage ratio, and can simultaneously achieve superior attack performance. Finally, the key message we want to convey through this paper is that: \textbf{it is reckless to consider guardrail moderation as a clutch at straws towards harmful fine-tuning attack}, as it cannot solve the inherent safety issue of the pre-trained LLMs. Our code is available at https://github.com/git-disl/Virus

Nitrogen Vacancy diamond nanoparticles (NVNPs) are increasingly integrated with methods for optical detection of magnetic resonance (ODMR), providing new opportunities in magnetic characterization that span the visualization of magnetic fields in microelectronic circuits, environmental sensing and biology. However, only a small number of studies utilize aggregates of NVNPs for surface-wide magnetometry being that spin orientations in aggregate NVNPs are inherently misaligned, precluding their use for proper magnetic field detection compared with expensive monocrystalline diamonds. A postprocessing method for layering NVNPs with aligned NV center orientations can potentially facilitate superior NV magnetometry by allowing sensitive detection combined with simplified probe preparation. We present a novel technology for creating densely stacked monolayers of NVNP with inherent interlayer alignment for sensitive measurement of local magnetic field perturbations in microelectronic traces. We establish spatial characteristics of deposited aggregates and demonstrate their ability to capture magnetic dipoles from conducting microwires via ODMR. Our approach forms a novel accessible protocol that can be used for broad applications in micromagnetometry.

Innovations in interaction design are increasingly driven by progress in machine learning fields. Automatic speech emotion recognition (SER) is such an example field on the rise, creating well performing models, which typically take as input a speech audio sample and provide as output digital labels or values describing the human emotion(s) embedded in the speech audio sample. Such labels and values are only abstract representations of the felt or expressed emotions, making it challenging to analyse them as experiences and work with them as design material for physical interactions, including tangible, haptic, or proxemic interactions. This paper argues that both the analysis of emotions and their use in interaction designs would benefit from alternative physical representations, which can be directly felt and socially communicated as bodily sensations or spatial behaviours. To this end, a method is described and a starter kit for speech emotion conversion is provided. Furthermore, opportunities of speech emotion conversion for new interaction designs are introduced, such as for interacting with animals or robots., Comment: 5 pages, 2 figures

Anthropogenic climate change has increased the probability, severity, and duration of heat waves and droughts, subsequently escalating the risk of wildfires. Mathematical and computational models can enhance our understanding of wildfire propagation dynamics. In this work, we present a simplified Advection-Diffusion-Reaction (ADR) model that accounts for the effect of fuel moisture, and also considers wind, local radiation, natural convection and topography. The model explicitly represents fuel moisture effects by means of the apparent calorific capacity method, distinguishing between live and dead fuel moisture content. Using this model, we conduct exploratory simulations and present theoretical insights into various modeling decisions in the context of ADR-based models. We aim to shed light on the interplay between the different modeled mechanisms in wildfire propagation to identify key factors influencing fire spread and to estimate the model's predictive capacity.

Alignment of pretrained LLMs using instruction-based datasets is critical for creating fine-tuned models that reflect human preference. A growing number of alignment-based fine-tuning algorithms and benchmarks emerged recently, fueling the efforts on effective alignments of pre-trained LLMs to ensure helpful, harmless, and honest answers from both open-source and closed-source LLMs. This paper tackles this problem by developing an alignment fusion approach, coined as $H^3$Fusion, with three unique characteristics. First, $H^3$Fusion ensembles multiple individually aligned LLMs to create a final fine-tuned alignment model with enhanced capabilities beyond those of individual models, delivering robust alignment through promoting helpful, harmless, honest fusion. Second, $H^3$Fusion leverages the mixture-of-experts (MoE) methodology in two steps. We first freeze the multi-head attention weights of each individual model while tuning the FFN layer during alignment fusion. Then we merge the aligned model weights with an expert router according to the type of input instruction and dynamically select a subset of experts that are best suited for producing the output response. Finally, we boost the performance of the resulting $H^3$3Fusion model by introducing gating loss and regularization terms. The former penalizes the selection errors of the expert-router, and the latter mediates the expert weights drifting during fine-tuning and dynamically adjusts the fusion behavior of the resulting model by canalizing the activations on the experts. Extensive evaluations on three benchmark datasets show that $H^3$3Fusion is more helpful, less harmful, and more honest from two aspects: it outperforms each individually aligned model by $11.37\%$, and it provides stronger robustness compared to the state-of-the-art LLM ensemble approaches by $13.77\%$. Code is available at github.com/sftekin/h3fusion.

The integration of Inverter-Based Resource (IBR) model into phasor-domain short circuit (SC) solvers challenges their numerical stability. To address the challenge, this paper proposes a solver that improves numerical stability by employing the Newton-Raphson iterative method. The solver can integrate the latest implementation of IBR SC model in industry-standard fault analysis programs including the voltage controlled current source tabular model as well as vendor-specific black-box and white-box equation-based models. The superior numerical stability of the proposed solver has been mathematically demonstrated, with identified convergence conditions. An algorithm for the implementation of the proposed solver in fault analysis programs has been developed. The objective is to improve the capability of the industry to accurately represent IBRs in SC studies and ensure system protection reliability in an IBR-dominated future.

As natural disasters become more frequent and severe, ensuring a resilient communications infrastructure is of paramount importance for effective disaster response and recovery. This disaster-resilient infrastructure should also respond to sustainability goals by providing an energy-efficient and economically feasible network that is accessible to everyone. This paper provides a comprehensive exploration of the technological solutions and strategies necessary to build and maintain resilient communications networks that can withstand and quickly recover from disaster scenarios. The paper starts with a survey of existing literature and related reviews to establish a solid foundation, followed by an overview of the global landscape of disaster communications and power supply management. We then introduce the key enablers of communications and energy resource technologies to support communications infrastructure, examining emerging trends that improve the resilience of these systems. Pre-disaster planning is emphasized as a critical phase where proactive communication and energy supply strategies can significantly mitigate the impact of disasters. We explore the essential technologies for disaster response, focusing on real-time communications and energy solutions that support rapid deployment and coordination in times of crisis. The paper presents post-disaster communication and energy management planning for effective rescue and evacuation operations. The main findings derived from the comprehensive survey are also summarized for each disaster phase. This is followed by an analysis of existing vendor products and services as well as standardization efforts and ongoing projects that contribute to the development of resilient infrastructures. A detailed case study of the Turkiye earthquakes is presented to illustrate the practical application of these technologies and strategies., Comment: 32 pages, submitted to IEEE

Combining large language models during training or at inference time has shown substantial performance gain over component LLMs. This paper presents LLM-TOPLA, a diversity-optimized LLM ensemble method with three unique properties: (i) We introduce the focal diversity metric to capture the diversity-performance correlation among component LLMs of an ensemble. (ii) We develop a diversity-optimized ensemble pruning algorithm to select the top-k sub-ensembles from a pool of $N$ base LLMs. Our pruning method recommends top-performing LLM subensembles of size $S$, often much smaller than $N$. (iii) We generate new output for each prompt query by utilizing a learn-to-ensemble approach, which learns to detect and resolve the output inconsistency among all component LLMs of an ensemble. Extensive evaluation on four different benchmarks shows good performance gain over the best LLM ensemble methods: (i) In constrained solution set problems, LLM-TOPLA outperforms the best-performing ensemble (Mixtral) by 2.2\% in accuracy on MMLU and the best-performing LLM ensemble (MoreAgent) on GSM8k by 2.1\%. (ii) In generative tasks, LLM-TOPLA outperforms the top-2 performers (Llama70b/Mixtral) on SearchQA by $3.9\mathrm{x}$ in F1, and on XSum by more than $38$ in ROUGE-1. Our code and dataset, which contains outputs of 8 modern LLMs on 4 benchmarks is available at https://github.com/git-disl/llm-topla

Recent research demonstrates that the nascent fine-tuning-as-a-service business model exposes serious safety concerns -- fine-tuning over a few harmful data uploaded by the users can compromise the safety alignment of the model. The attack, known as harmful fine-tuning attack, has raised a broad research interest among the community. However, as the attack is still new, \textbf{we observe that there are general misunderstandings within the research community.} To clear up concern, this paper provide a comprehensive overview to three aspects of harmful fine-tuning: attacks setting, defense design and evaluation methodology. Specifically, we first present the threat model of the problem, and introduce the harmful fine-tuning attack and its variants. Then we systematically survey the existing literature on attacks/defenses/mechanical analysis of the problem. Finally, we introduce the evaluation methodology and outline future research directions that might contribute to the development of the field. Additionally, we present a list of questions of interest, which might be useful to refer to when reviewers in the peer review process question the realism of the experiment/attack/defense setting. A curated list of relevant papers is maintained and made accessible at: https://github.com/git-disl/awesome_LLM-harmful-fine-tuning-papers.

For users navigating travel e-commerce websites, the process of researching products and making a purchase often results in intricate browsing patterns that span numerous sessions over an extended period of time. The resulting clickstream data chronicle these user journeys and present valuable opportunities to derive insights that can significantly enhance personalized recommendations. We introduce TRACE, a novel transformer-based approach tailored to generate rich user embeddings from live multi-session clickstreams for real-time recommendation applications. Prior works largely focus on single-session product sequences, whereas TRACE leverages site-wide page view sequences spanning multiple user sessions to model long-term engagement. Employing a multi-task learning framework, TRACE captures comprehensive user preferences and intents distilled into low-dimensional representations. We demonstrate TRACE's superior performance over vanilla transformer and LLM-style architectures through extensive experiments on a large-scale travel e-commerce dataset of real user journeys, where the challenges of long page-histories and sparse targets are particularly prevalent. Visualizations of the learned embeddings reveal meaningful clusters corresponding to latent user states and behaviors, highlighting TRACE's potential to enhance recommendation systems by capturing nuanced user interactions and preferences, Comment: RecSys Workshop on Recommenders in Tourism (RecTour 2024), October 14th-18th, 2024, co-located with the 18th ACM Conference on Recommender Systems, Bari, Italy

Harmful fine-tuning issue \citep{qi2023fine} poses serious safety concerns for Large language models' fine-tuning-as-a-service. While existing defenses \citep{huang2024vaccine,rosati2024representation} have been proposed to mitigate the issue, their performances are still far away from satisfactory, and the root cause of the problem has not been fully recovered. For the first time in the literature, we in this paper show that \textit{harmful perturbation} over the model weights should be the root cause of alignment-broken of harmful fine-tuning. In order to attenuate the negative impact of harmful perturbation, we propose an alignment-stage solution, dubbed Booster. Technically, along with the original alignment loss, we append a loss regularizer in the alignment stage's optimization. The regularizer ensures that the model's harmful loss reduction before/after simulated harmful perturbation is attenuated, thereby mitigating the subsequent fine-tuning risk. Empirical results show that Booster can effectively reduce the harmful score of the fine-tuned models while maintaining the performance of downstream tasks. Our code is available at \url{https://github.com/git-disl/Booster}.

Neural network models for audio tasks, such as automatic speech recognition (ASR) and acoustic scene classification (ASC), are susceptible to noise contamination for real-life applications. To improve audio quality, an enhancement module, which can be developed independently, is explicitly used at the front-end of the target audio applications. In this paper, we present an end-to-end learning solution to jointly optimise the models for audio enhancement (AE) and the subsequent applications. To guide the optimisation of the AE module towards a target application, and especially to overcome difficult samples, we make use of the sample-wise performance measure as an indication of sample importance. In experiments, we consider four representative applications to evaluate our training paradigm, i.e., ASR, speech command recognition (SCR), speech emotion recognition (SER), and ASC. These applications are associated with speech and non-speech tasks concerning semantic and non-semantic features, transient and global information, and the experimental results indicate that our proposed approach can considerably boost the noise robustness of the models, especially at low signal-to-noise ratios (SNRs), for a wide range of computer audition tasks in everyday-life noisy environments.

Telling stories is an integral part of human communication which can evoke emotions and influence the affective states of the audience. Automatically modeling emotional trajectories in stories has thus attracted considerable scholarly interest. However, as most existing works have been limited to unsupervised dictionary-based approaches, there is no benchmark for this task. We address this gap by introducing continuous valence and arousal labels for an existing dataset of children's stories originally annotated with discrete emotion categories. We collect additional annotations for this data and map the categorical labels to the continuous valence and arousal space. For predicting the thus obtained emotionality signals, we fine-tune a DeBERTa model and improve upon this baseline via a weakly supervised learning approach. The best configuration achieves a Concordance Correlation Coefficient (CCC) of $.8221$ for valence and $.7125$ for arousal on the test set, demonstrating the efficacy of our proposed approach. A detailed analysis shows the extent to which the results vary depending on factors such as the author, the individual story, or the section within the story. In addition, we uncover the weaknesses of our approach by investigating examples that prove to be difficult to predict., Comment: Accepted to ACL 2024 Findings. arXiv admin note: text overlap with arXiv:2212.11382

Recent studies show that Large Language Models (LLMs) with safety alignment can be jail-broken by fine-tuning on a dataset mixed with harmful data. First time in the literature, we show that the jail-broken effect can be mitigated by separating states in the finetuning stage to optimize the alignment and user datasets. Unfortunately, our subsequent study shows that this simple Bi-State Optimization (BSO) solution experiences convergence instability when steps invested in its alignment state is too small, leading to downgraded alignment performance. By statistical analysis, we show that the \textit{excess drift} towards consensus could be a probable reason for the instability. To remedy this issue, we propose \textbf{L}azy(\textbf{i}) \textbf{s}afety \textbf{a}lignment (\textbf{Lisa}), which introduces a proximal term to constraint the drift of each state. Theoretically, the benefit of the proximal term is supported by the convergence analysis, wherein we show that a sufficient large proximal factor is necessary to guarantee Lisa's convergence. Empirically, our results on four downstream finetuning tasks show that Lisa with a proximal term can significantly increase alignment performance while maintaining the LLM's accuracy on the user tasks. Code is available at \url{https://github.com/git-disl/Lisa}., Comment: Accepted by NeurIPS2024. arXiv admin note: substantial text overlap with arXiv:2402.01109

Natural disasters can have catastrophic consequences, a poignant example is the series of $7.7$ and $7.6$ magnitude earthquakes that devastated T\"urkiye on February 6, 2023. To limit the damage, it is essential to maintain the communications infrastructure to ensure individuals impacted by the disaster can receive critical information. The disastrous earthquakes in T\"urkiye have revealed the importance of considering communications and energy solutions together to build resilient and sustainable infrastructure. Thus, this paper proposes an integrated space-air-ground-sea network architecture that utilizes various communications and energy-enabling technologies. This study aims to contribute to the development of robust and sustainable disaster-response frameworks. In light of the T\"urkiye earthquakes, two methods for network management are proposed: the first aims to ensure sustainability in the pre-disaster phase and the second aims to maintain communications during the in-disaster phase. In these frameworks, communications technologies such as High Altitude Platform Station(s)(HAPS), which are among the key enablers to unlock the potential of 6G networks, and energy technologies such as Renewable Energy Sources (RES), Battery Energy Storage Systems (BESSs), and Electric Vehicles (EVs) have been used as the prominent technologies. By simulating a case study, we demonstrate the performance of a proposed framework for providing network resiliency. The paper concludes with potential challenges and future directions to achieve a disaster-resilient network architecture solution., Comment: accepted in IEEE COMMAG

This paper presents FusionShot, a focal diversity optimized few-shot ensemble learning approach for boosting the robustness and generalization performance of pre-trained few-shot models. The paper makes three original contributions. First, we explore the unique characteristics of few-shot learning to ensemble multiple few-shot (FS) models by creating three alternative fusion channels. Second, we introduce the concept of focal error diversity to learn the most efficient ensemble teaming strategy, rather than assuming that an ensemble of a larger number of base models will outperform those sub-ensembles of smaller size. We develop a focal-diversity ensemble pruning method to effectively prune out the candidate ensembles with low ensemble error diversity and recommend top-$K$ FS ensembles with the highest focal error diversity. Finally, we capture the complex non-linear patterns of ensemble few-shot predictions by designing the learn-to-combine algorithm, which can learn the diverse weight assignments for robust ensemble fusion over different member models. Extensive experiments on representative few-shot benchmarks show that the top-K ensembles recommended by FusionShot can outperform the representative SOTA few-shot models on novel tasks (different distributions and unknown at training), and can prevail over existing few-shot learners in both cross-domain settings and adversarial settings. For reproducibility purposes, FusionShot trained models, results, and code are made available at https://github.com/sftekin/fusionshot