An explainable transformer circuit for compositional generalization

Compositional generalization-the systematic combination of known components into novel structures-remains a core challenge in cognitive science and machine learning. Although transformer-based large language models can exhibit strong performance on certain compositional tasks, the underlying mechanisms driving these abilities remain opaque, calling into question their interpretability. In this work, we identify and mechanistically interpret the circuit responsible for compositional induction in a compact transformer. Using causal ablations, we validate the circuit and formalize its operation using a program-like description. We further demonstrate that this mechanistic understanding enables precise activation edits to steer the model's behavior predictably. Our findings advance the understanding of complex behaviors in transformers and highlight such insights can provide a direct pathway for model control.

LLaVA-SLT: Visual Language Tuning for Sign Language Translation

In the realm of Sign Language Translation (SLT), reliance on costly gloss-annotated datasets has posed a significant barrier. Recent advancements in gloss-free SLT methods have shown promise, yet they often largely lag behind gloss-based approaches in terms of translation accuracy. To narrow this performance gap, we introduce LLaVA-SLT, a pioneering Large Multimodal Model (LMM) framework designed to leverage the power of Large Language Models (LLMs) through effectively learned visual language embeddings. Our model is trained through a trilogy. First, we propose linguistic continued pretraining. We scale up the LLM and adapt it to the sign language domain using an extensive corpus dataset, effectively enhancing its textual linguistic knowledge about sign language. Then, we adopt visual contrastive pretraining to align the visual encoder with a large-scale pretrained text encoder. We propose hierarchical visual encoder that learns a robust word-level intermediate representation that is compatible with LLM token embeddings. Finally, we propose visual language tuning. We freeze pretrained models and employ a lightweight trainable MLP connector. It efficiently maps the pretrained visual language embeddings into the LLM token embedding space, enabling downstream SLT task. Our comprehensive experiments demonstrate that LLaVA-SLT outperforms the state-of-the-art methods. By using extra annotation-free data, it even closes to the gloss-based accuracy.

Knowledge Distillation in RNN-Attention Models for Early Prediction of Student Performance

Educational data mining (EDM) is a part of applied computing that focuses on automatically analyzing data from learning contexts. Early prediction for identifying at-risk students is a crucial and widely researched topic in EDM research. It enables instructors to support at-risk students to stay on track, preventing student dropout or failure. Previous studies have predicted students' learning performance to identify at-risk students by using machine learning on data collected from e-learning platforms. However, most studies aimed to identify at-risk students utilizing the entire course data after the course finished. This does not correspond to the real-world scenario that at-risk students may drop out before the course ends. To address this problem, we introduce an RNN-Attention-KD (knowledge distillation) framework to predict at-risk students early throughout a course. It leverages the strengths of Recurrent Neural Networks (RNNs) in handling time-sequence data to predict students' performance at each time step and employs an attention mechanism to focus on relevant time steps for improved predictive accuracy. At the same time, KD is applied to compress the time steps to facilitate early prediction. In an empirical evaluation, RNN-Attention-KD outperforms traditional neural network models in terms of recall and F1-measure. For example, it obtained recall and F1-measure of 0.49 and 0.51 for Weeks 1--3 and 0.51 and 0.61 for Weeks 1--6 across all datasets from four years of a university course. Then, an ablation study investigated the contributions of different knowledge transfer methods (distillation objectives). We found that hint loss from the hidden layer of RNN and context vector loss from the attention module on RNN could enhance the model's prediction performance for identifying at-risk students. These results are relevant for EDM researchers employing deep learning models.

Attention-Seeker: Dynamic Self-Attention Scoring for Unsupervised Keyphrase Extraction

This paper proposes Attention-Seeker, an unsupervised keyphrase extraction method that leverages self-attention maps from a Large Language Model to estimate the importance of candidate phrases. Our approach identifies specific components - such as layers, heads, and attention vectors - where the model pays significant attention to the key topics of the text. The attention weights provided by these components are then used to score the candidate phrases. Unlike previous models that require manual tuning of parameters (e.g., selection of heads, prompts, hyperparameters), Attention-Seeker dynamically adapts to the input text without any manual adjustments, enhancing its practical applicability. We evaluate Attention-Seeker on four publicly available datasets: Inspec, SemEval2010, SemEval2017, and Krapivin. Our results demonstrate that, even without parameter tuning, Attention-Seeker outperforms most baseline models, achieving state-of-the-art performance on three out of four datasets, particularly excelling in extracting keyphrases from long documents.

On the Feasibility of Fingerprinting Collaborative Robot Traffic

This study examines privacy risks in collaborative robotics, focusing on the potential for traffic analysis in encrypted robot communications. While previous research has explored low-level command recovery, our work investigates high-level motion recovery from command message sequences. We evaluate the efficacy of traditional website fingerprinting techniques (k-FP, KNN, and CUMUL) and their limitations in accurately identifying robotic actions due to their inability to capture detailed temporal relationships. To address this, we introduce a traffic classification approach using signal processing techniques, demonstrating high accuracy in action identification and highlighting the vulnerability of encrypted communications to privacy breaches. Additionally, we explore defenses such as packet padding and timing manipulation, revealing the challenges in balancing traffic analysis resistance with network efficiency. Our findings emphasize the need for continued development of practical defenses in robotic privacy and security.

RoboSync: OS for Social Robots with Customizable Behaviour

Traditional robotic systems require complex implementations that are not always accessible or easy to use for Human-Robot Interaction (HRI) application developers. With the aim of simplifying the implementation of HRI applications, this paper introduces a novel real-time operating system (RTOS) designed for customizable HRI - RoboSync. By creating multi-level abstraction layers, the system enables users to define complex emotional and behavioral models without needing deep technical expertise. The system's modular architecture comprises a behavior modeling layer, a machine learning plugin configuration layer, a sensor checks customization layer, a scheduler that fits the need of HRI, and a communication and synchronization layer. This approach not only promotes ease of use without highly specialized skills but also ensures real-time responsiveness and adaptability. The primary functionality of the RTOS has been implemented for proof of concept and was tested on a CortexM4 microcontroller, demonstrating its potential for a wide range of lightweight simple-to-implement social robotics applications.

Augmented Kinesthetic Teaching: Enhancing Task Execution Efficiency through Intuitive Human Instructions

In this paper, we present a complete and efficient implementation of a knowledge-sharing augmented kinesthetic teaching approach for efficient task execution in robotics. Our augmented kinesthetic teaching method integrates intuitive human feedback, including verbal, gesture, gaze, and physical guidance, to facilitate the extraction of multiple layers of task information including control type, attention direction, input and output type, action state change trigger, etc., enhancing the adaptability and autonomy of robots during task execution. We propose an efficient Programming by Demonstration (PbD) framework for users with limited technical experience to teach the robot in an intuitive manner. The proposed framework provides an interface for such users to teach customized tasks using high-level commands, with the goal of achieving a smoother teaching experience and task execution. This is demonstrated with the sample task of pouring water.

On the tightness of linear relaxation based robustness certification methods

There has been a rapid development and interest in adversarial training and defenses in the machine learning community in the recent years. One line of research focuses on improving the performance and efficiency of adversarial robustness certificates for neural networks \cite{gowal:19, wong_zico:18, raghunathan:18, WengTowardsFC:18, wong:scalable:18, singh:convex_barrier:19, Huang_etal:19, single-neuron-relax:20, Zhang2020TowardsSA}. While each providing a certification to lower (or upper) bound the true distortion under adversarial attacks via relaxation, less studied was the tightness of relaxation. In this paper, we analyze a family of linear outer approximation based certificate methods via a meta algorithm, IBP-Lin. The aforementioned works often lack quantitative analysis to answer questions such as how does the performance of the certificate method depend on the network configuration and the choice of approximation parameters. Under our framework, we make a first attempt at answering these questions, which reveals that the tightness of linear approximation based certification can depend heavily on the configuration of the trained networks.

Neural document expansion for ad-hoc information retrieval

Recently, Nogueira et al. [2019] proposed a new approach to document expansion based on a neural Seq2Seq model, showing significant improvement on short text retrieval task. However, this approach needs a large amount of in-domain training data. In this paper, we show that this neural document expansion approach can be effectively adapted to standard IR tasks, where labels are scarce and many long documents are present.

Exponentially convergent stochastic k-PCA without variance reduction

We present Matrix Krasulina, an algorithm for online k-PCA, by generalizing the classic Krasulina's method (Krasulina, 1969) from vector to matrix case. We show, both theoretically and empirically, that the algorithm naturally adapts to data low-rankness and converges exponentially fast to the ground-truth principal subspace. Notably, our result suggests that despite various recent efforts to accelerate the convergence of stochastic-gradient based methods by adding a O(n)-time variance reduction step, for the k-PCA problem, a truly online SGD variant suffices to achieve exponential convergence on intrinsically low-rank data.