Search

Recent large language models (LLMs) have excelled across a wide range of tasks, but their use in high-stakes and compute-limited settings has intensified the demand for interpretability and efficiency. We address this need by proposing Induction-head ngram models (Induction-Gram), a method that builds an efficient, interpretable LM by bolstering modern ngram models with a hand-engineered "induction head". This induction head uses a custom neural similarity metric to efficiently search the model's input context for potential next-word completions. This process enables Induction-Gram to provide ngram-level grounding for each generated token. Moreover, experiments show that this simple method significantly improves next-word prediction over baseline interpretable models (up to 26%p) and can be used to speed up LLM inference for large models through speculative decoding. We further study Induction-Gram in a natural-language neuroscience setting, where the goal is to predict the next fMRI response in a sequence. It again provides a significant improvement over interpretable models (20% relative increase in the correlation of predicted fMRI responses), potentially enabling deeper scientific investigation of language selectivity in the brain. The code is available at https://github.com/ejkim47/induction-gram.

Recent research has shown that Large Language Models (LLMs) are vulnerable to automated jailbreak attacks, where adversarial suffixes crafted by algorithms appended to harmful queries bypass safety alignment and trigger unintended responses. Current methods for generating these suffixes are computationally expensive and have low Attack Success Rates (ASR), especially against well-aligned models like Llama2 and Llama3. To overcome these limitations, we introduce ADV-LLM, an iterative self-tuning process that crafts adversarial LLMs with enhanced jailbreak ability. Our framework significantly reduces the computational cost of generating adversarial suffixes while achieving nearly 100\% ASR on various open-source LLMs. Moreover, it exhibits strong attack transferability to closed-source models, achieving 99% ASR on GPT-3.5 and 49% ASR on GPT-4, despite being optimized solely on Llama3. Beyond improving jailbreak ability, ADV-LLM provides valuable insights for future safety alignment research through its ability to generate large datasets for studying LLM safety., Comment: 18 pages

Power system operators must ensure that dispatch decisions remain feasible in case of grid outages or contingencies to prevent cascading failures and ensure reliable operation. However, checking the feasibility of all $N - k$ contingencies -- every possible simultaneous failure of $k$ grid components -- is computationally intractable for even small $k$, requiring system operators to resort to heuristic screening methods. Because of the increase in uncertainty and changes in system behaviors, heuristic lists might not include all relevant contingencies, generating false negatives in which unsafe scenarios are misclassified as safe. In this work, we propose to use input-convex neural networks (ICNNs) for contingency screening. We show that ICNN reliability can be determined by solving a convex optimization problem, and by scaling model weights using this problem as a differentiable optimization layer during training, we can learn an ICNN classifier that is both data-driven and has provably guaranteed reliability. Namely, our method can ensure a zero false negative rate. We empirically validate this methodology in a case study on the IEEE 39-bus test network, observing that it yields substantial (10-20x) speedups while having excellent classification accuracy., Comment: 11 pages, 4 figures

Artificial intelligence empowered autonomous vehicles and robotics have to sense the fast-changing three-dimensional environment with high precision and speed. However, it is challenging for the state-of-the-art ambiguity-free light detection and ranging (LiDAR) techniques to achieve absolute distance measurement with simultaneous high precision and high acquisition rate. Here we demonstrate an electro-optic frequency comb-enabled precise absolute distance measurement method, repetition rate modulated frequency comb (RRMFC), with megahertz-level acquisition rate. To achieve RRMFC, we designed and fabricated an integrated lithium niobate phase modulator with a modulation length of 5 cm and a half-wave voltage of 1.52 V, leading to over 50 sidebands and a continuously tunable repetition rate. Leveraging these unique features, RRMFC can directly resolve distance in time domain, leading to an acquisition rate as high as 25 MHz and an Allan deviation down to 13.77 {\mu}m at an averaging time of 724 {\mu}s. Based on RRMFC, we achieved high-speed 3D imaging at millimeter-level precision with a single laser. RRMFC-based LiDAR allows the autonomous vehicles and robotics to sense the fine details of fast-changing environment with high precision., Comment: These authors contributed equally: Yifan Qi, Xingyu Jia

Binary code summarization, while invaluable for understanding code semantics, is challenging due to its labor-intensive nature. This study delves into the potential of large language models (LLMs) for binary code comprehension. To this end, we present BinSum, a comprehensive benchmark and dataset of over 557K binary functions and introduce a novel method for prompt synthesis and optimization. To more accurately gauge LLM performance, we also propose a new semantic similarity metric that surpasses traditional exact-match approaches. Our extensive evaluation of prominent LLMs, including ChatGPT, GPT-4, Llama 2, and Code Llama, reveals 10 pivotal insights. This evaluation generates 4 billion inference tokens, incurred a total expense of 11,418 US dollars and 873 NVIDIA A100 GPU hours. Our findings highlight both the transformative potential of LLMs in this field and the challenges yet to be overcome.

The emergence of human-like abilities of AI systems for content generation in domains such as text, audio, and vision has prompted the development of classifiers to determine whether content originated from a human or a machine. Implicit in these efforts is an assumption that the generation properties of a human are different from that of the machine. In this work, we provide a framework in the language of statistical pattern recognition that quantifies the difference between the distributions of human and machine-generated content conditioned on an evaluation context. We describe current methods in the context of the framework and demonstrate how to use the framework to evaluate the progression of generative models towards human-like capabilities, among many axes of analysis.

Energy storage (ES) and virtual energy storage (VES) are key components to realizing power system decarbonization. Although ES and VES have been proven to deliver various types of grid services, little work has so far provided a systematical framework for quantifying their adequacy contribution and credible capacity value while incorporating human and market behavior. Therefore, this manuscript proposed a novel evaluation framework to evaluate the capacity credit (CC) of ES and VES. To address the system capacity inadequacy and market behavior of storage, a two-stage coordinated dispatch is proposed to achieve the trade-off between day-ahead self-energy management of resources and efficient adjustment to real-time failures. And we further modeled the human behavior with storage operations and incorporate two types of decision-independent uncertainties (DIUs) (operate state and self-consumption) and one type of decision-dependent uncertainty (DDUs) (available capacity) into the proposed dispatch. Furthermore, novel reliability and CC indices (e.g., equivalent physical storage capacity (EPSC)) are introduced to evaluate the practical and theoretical adequacy contribution of ES and VES, as well as the ability to displace generation and physical storage while maintaining equivalent system adequacy. Exhaustive case studies based on the IEEE RTS-79 system and real-world data verify the significant consequence (10%-70% overestimated CC) of overlooking DIUs and DDUs in the previous works, while the proposed method outperforms other and can generate a credible and realistic result. Finally, we investigate key factors affecting the adequacy contribution of ES and VES, and reasonable suggestions are provided for better flexibility utilization of ES and VES in decarbonized power system., Comment: capacity credit, decision-dependent uncertainty, decarbonized power system

Recent advancements have been made in the development of cell-based in-vitro neuronal networks, or organoids. In order to better understand the network structure of these organoids, a super-selective algorithm has been proposed for inferring the effective connectivity networks from multi-electrode array data. In this paper, we apply a novel statistical method called spectral mirror estimation to the time series of inferred effective connectivity organoid networks. This method produces a one-dimensional iso-mirror representation of the dynamics of the time series of the networks which exhibits a piecewise linear structure. A classical change point algorithm is then applied to this representation, which successfully detects a change point coinciding with the neuroscientifically significant time inhibitory neurons start appearing and the percentage of astrocytes increases dramatically. This finding demonstrates the potential utility of applying the iso-mirror dynamic structure discovery method to inferred effective connectivity time series of organoid networks.

We propose a class of models based on Fisher's Linear Discriminant (FLD) in the context of domain adaptation. The class is the convex combination of two hypotheses: i) an average hypothesis representing previously seen source tasks and ii) a hypothesis trained on a new target task. For a particular generative setting we derive the optimal convex combination of the two models under 0-1 loss, propose a computable approximation, and study the effect of various parameter settings on the relative risks between the optimal hypothesis, hypothesis i), and hypothesis ii). We demonstrate the effectiveness of the proposed optimal classifier in the context of EEG- and ECG-based classification settings and argue that the optimal classifier can be computed without access to direct information from any of the individual source tasks. We conclude by discussing further applications, limitations, and possible future directions.

A key barrier to using reinforcement learning (RL) in many real-world applications is the requirement of a large number of system interactions to learn a good control policy. Off-policy and Offline RL methods have been proposed to reduce the number of interactions with the physical environment by learning control policies from historical data. However, their performances suffer from the lack of exploration and the distributional shifts in trajectories once controllers are updated. Moreover, most RL methods require that all states are directly observed, which is difficult to be attained in many settings. To overcome these challenges, we propose a trajectory generation algorithm, which adaptively generates new trajectories as if the system is being operated and explored under the updated control policies. Motivated by the fundamental lemma for linear systems, assuming sufficient excitation, we generate trajectories from linear combinations of historical trajectories. For linear feedback control, we prove that the algorithm generates trajectories with the exact distribution as if they are sampled from the real system using the updated control policy. In particular, the algorithm extends to systems where the states are not directly observed. Experiments show that the proposed method significantly reduces the number of sampled data needed for RL algorithms.

In this paper, we study an unmanned aerial vehicle (UAV) enabled data collection system, where an intelligent reflecting surface (IRS) is deployed to assist in the communication from a cluster of Internet-of-Things (IoT) devices to a UAV in the presence of a jammer. We aim to improve the energy efficiency (EE) via the joint design of UAV trajectory, IRS passive beamforming, device power allocation, and communication scheduling. However, the formulated non-linear fractional programming problem is challenging to solve due to its non-convexity and coupled variables. To overcome the difficulty, we propose an alternating optimization based algorithm to solve it sub-optimally by leveraging Dinkelbach's algorithm, successive convex approximation (SCA) technique, and block coordinate descent (BCD) method. Extensive simulation results show that the proposed design can significantly improve the anti-jamming performance. In particular, for the remote jammer case, the proposed design can largely shorten the flight path and thus decrease the energy consumption via the signal enhancement; while for the local jammer case, which is deemed highly challenging in conventional systems without IRS since the retreating away strategy becomes ineffective, our proposed design even achieves a higher performance gain owing to the efficient jamming signal mitigation., Comment: Exploiting IRS for reducing energy consumption and shortening flight paths in UAV communications facing malicious jamming

Deep neural networks are susceptible to label noise. Existing methods to improve robustness, such as meta-learning and regularization, usually require significant change to the network architecture or careful tuning of the optimization procedure. In this work, we propose a simple hierarchical approach that incorporates a label hierarchy when training the deep learning models. Our approach requires no change of the network architecture or the optimization procedure. We investigate our hierarchical network through a wide range of simulated and real datasets and various label noise types. Our hierarchical approach improves upon regular deep neural networks in learning with label noise. Combining our hierarchical approach with pre-trained models achieves state-of-the-art performance in real-world noisy datasets., Comment: 8 pages, 7 figures