We are unable to provide the monkey data as .

Large Language Models (LLMs) are trained to support an increasing number of languages, yet their predefined tokenizers remain a bottleneck for adapting models to lower-resource or distinct-script languages. Existing tokenizer transfer methods typically rely on semantic heuristics to initialize new embeddings, ignoring higher-layer model dynamics and limiting transfer quality. We propose Model-A ware Tokenizer Transfer (MA TT), a method that incorporates model internals into the tokenizer transfer process. MA TT introduces an Attention Influence Modeling (AIM) objective that distills inter-token communication patterns from a source model into a target model with a new tokenizer, providing an efficient warm-up before standard language modeling. Unlike approaches that focus solely on embedding similarity, MA TT leverages attention behavior to guide embedding initialization and adaptation. Experiments across diverse linguistic settings show that MA TT recovers a large fraction of the original model's performance within a few GPU hours, outperforming heuristic baselines. These results demonstrate that incorporating model-level signals offers a practical and effective path toward robust tokenizer transfer in multilingual LLMs. Recent advances in large language models (LLMs) have shifted attention from training monolingual models (Jiang et al., 2023; Touvron et al., 2023) to covering an increasing number of languages (Grattafiori et al., 2024; Team et al., 2025).

In general, we have found the JSLDS loss function strengths to be relatively easy to select (see example settings in the specific experiment sections below). RNN's fixed points or slow points would defeat the primary purpose of the method. However, other variations are possible. We set the number of timesteps T = 25. We trained both methods with the Adam optimizer with default settings.

While software requirements are often expressed in natural language, verifying the correctness of a program against natural language requirements is a hard and underexplored problem. Large language models (LLMs) are promising candidates for addressing this challenge, however our experience shows that they are ineffective in this task, often failing to detect even straightforward bugs. To address this gap, we introduce HoarePrompt, a novel approach that adapts fundamental ideas from program analysis and verification to natural language artifacts. Drawing inspiration from the strongest postcondition calculus, HoarePrompt employs a systematic, step-by-step process in which an LLM generates natural language descriptions of reachable program states at various points in the code. To manage loops, we propose few-shot-driven k-induction, an adaptation of the k-induction method widely used in model checking. Once program states are described, HoarePrompt leverages the LLM to assess whether the program, annotated with these state descriptions, conforms to the natural language requirements. For evaluating the quality of classifiers of program correctness with respect to natural language requirements, we constructed CoCoClaNeL, a challenging dataset of solutions to programming competition problems. Our experiments show that HoarePrompt improves the MCC by 62% compared to directly using Zero-shot-CoT prompts for correctness classification. Furthermore, HoarePrompt outperforms a classifier that assesses correctness via LLM-based test generation by increasing the MCC by 93%. The inductive reasoning mechanism contributes a 28% boost to MCC, underscoring its effectiveness in managing loops.

Crossbar memory arrays have been touted as the workhorse of in-memory computing (IMC)-based acceleration of Deep Neural Networks (DNNs), but the associated hardware non-idealities limit their efficacy. To address this, cross-layer design solutions that reduce the impact of hardware non-idealities on DNN accuracy are needed. In Part 1 of this paper, we established the co-optimization strategies for various memory technologies and their crossbar arrays, and conducted a comparative technology evaluation in the context of IMC robustness. In this part, we analyze various design knobs such as array size and bit-slice (number of bits per device) and their impact on the performance of 8T SRAM, ferroelectric transistor (FeFET), Resistive RAM (ReRAM) and spin-orbit-torque magnetic RAM (SOT-MRAM) in the context of inference accuracy at 7nm technology node. Further, we study the effect of circuit design solutions such as Partial Wordline Activation (PWA) and custom ADC reference levels that reduce the hardware non-idealities and comparatively analyze the response of each technology to such accuracy enhancing techniques. Our results on ResNet-20 (with CIFAR-10) show that PWA increases accuracy by up to 32.56% while custom ADC reference levels yield up to 31.62% accuracy enhancement. We observe that compared to the other technologies, FeFET, by virtue of its small layout height and high distinguishability of its memory states, is best suited for large arrays. For higher bit-slices and a more complex dataset (ResNet-50 with Cifar-100) we found that ReRAM matches the performance of FeFET.

Quantum computers have the potential to outperform classical computers for some complex computational problems. However, current quantum computers (e.g., from IBM and Google) have inherent noise that results in errors in the outputs of quantum software executing on the quantum computers, affecting the reliability of quantum software development. The industry is increasingly interested in machine learning (ML)--based error mitigation techniques, given their scalability and practicality. However, existing ML-based techniques have limitations, such as only targeting specific noise types or specific quantum circuits. This paper proposes a practical ML-based approach, called Q-LEAR, with a novel feature set, to mitigate noise errors in quantum software outputs. We evaluated Q-LEAR on eight quantum computers and their corresponding noisy simulators, all from IBM, and compared Q-LEAR with a state-of-the-art ML-based approach taken as baseline. Results show that, compared to the baseline, Q-LEAR achieved a 25% average improvement in error mitigation on both real quantum computers and simulators. We also discuss the implications and practicality of Q-LEAR, which, we believe, is valuable for practitioners.

Generalization bounds are a critical tool to assess the training data requirements of Quantum Machine Learning (QML). Recent work has established guarantees for in-distribution generalization of quantum neural networks (QNNs), where training and testing data are drawn from the same data distribution. However, there are currently no results on out-of-distribution generalization in QML, where we require a trained model to perform well even on data drawn from a different distribution to the training distribution. Here, we prove out-of-distribution generalization for the task of learning an unknown unitary. In particular, we show that one can learn the action of a unitary on entangled states having trained only product states. Since product states can be prepared using only single-qubit gates, this advances the prospects of learning quantum dynamics on near term quantum hardware, and further opens up new methods for both the classical and quantum compilation of quantum circuits.

Recurrent neural networks play an important role in both research and industry. With the advent of quantum machine learning, the quantisation of recurrent neural networks has become recently relevant. We propose fully quantum recurrent neural networks, based on dissipative quantum neural networks, capable of learning general causal quantum automata. A quantum training algorithm is proposed and classical simulations for the case of product outputs with the fidelity as cost function are carried out. We thereby demonstrate the potential of these algorithms to learn complex quantum processes with memory in terms of the exemplary delay channel, the time evolution of quantum states governed by a time-dependent Hamiltonian, and high- and low-frequency noise mitigation. Numerical simulations indicate that our quantum recurrent neural networks exhibit a striking ability to generalise from small training sets.

In the problem of quantum channel discrimination, one distinguishes between a given number of quantum channels, which is done by sending an input state through a channel and measuring the output state. This work studies applications of variational quantum circuits and machine learning techniques for discriminating such channels. In particular, we explore (i) the practical implementation of embedding this task into the framework of variational quantum computing, (ii) training a quantum classifier based on variational quantum circuits, and (iii) applying the quantum kernel estimation technique. For testing these three channel discrimination approaches, we considered a pair of entanglement-breaking channels and the depolarizing channel with two different depolarization factors. For the approach (i), we address solving the quantum channel discrimination problem using widely discussed parallel and sequential strategies. We show the advantage of the latter in terms of better convergence with less quantum resources. Quantum channel discrimination with a variational quantum classifier (ii) allows one to operate even with random and mixed input states and simple variational circuits. The kernel-based classification approach (iii) is also found effective as it allows one to discriminate depolarizing channels associated not with just fixed values of the depolarization factor, but with ranges of it. Additionally, we discovered that a simple modification of one of the commonly used kernels significantly increases the efficiency of this approach. Finally, our numerical findings reveal that the performance of variational methods of channel discrimination depends on the trace of the product of the output states. These findings demonstrate that quantum machine learning can be used to discriminate channels, such as those representing physical noise processes.

Cryptanalysis on standard quantum cryptographic systems generally involves finding optimal adversarial attack strategies on the underlying protocols. The core principle of modelling quantum attacks in many cases reduces to the adversary's ability to clone unknown quantum states which facilitates the extraction of some meaningful secret information. Explicit optimal attack strategies typically require high computational resources due to large circuit depths or, in many cases, are unknown. In this work, we propose variational quantum cloning (VQC), a quantum machine learning based cryptanalysis algorithm which allows an adversary to obtain optimal (approximate) cloning strategies with short depth quantum circuits, trained using hybrid classical-quantum techniques. The algorithm contains operationally meaningful cost functions with theoretical guarantees, quantum circuit structure learning and gradient descent based optimisation. Our approach enables the end-to-end discovery of hardware efficient quantum circuits to clone specific families of quantum states, which in turn leads to an improvement in cloning fidelites when implemented on quantum hardware: the Rigetti Aspen chip. Finally, we connect these results to quantum cryptographic primitives, in particular quantum coin flipping. We derive attacks on two protocols as examples, based on quantum cloning and facilitated by VQC. As a result, our algorithm can improve near term attacks on these protocols, using approximate quantum cloning as a resource.