paperlist.ai

AI conference papers

AI conference paper library

Search papers from 30+ top conferences, including NeurIPS, ICLR, ICML, CVPR, ACL and AAAI. Filter by keyword, then open any paper for its abstract, authors and links.

arXiv

MARS: Margin and Semantic-Aware Data Augmentation for Reward Modeling

Payel Bhattacharjee, Osvaldo Simeone, Ravi Tandon

Reward modeling is central to RLHF, RLAIF, and PPO-based alignment, but its reliability is often limited by scarce and heterogeneous human preference data. In this paper, we introduce MARS (Margin and Semantic-Aware Data Augmentation for Reward Modeling), an adaptive augmentation framework for controlled low-resource reward modeling. MARS allocates more augmentation to low-margin preference pairs and uses semantic-distance-based refinement to improve chosen-rejected contrast before generating synthetic preference samples. Across three preference datasets, two reward-model backbones, and downstream alignment evaluations, MARS improves average RewardBench performance and alignment win rates over uniform augmentation, WoN, and AdaBoost-style baselines. Ablations and independent-judge evaluations suggest that the gains are not solely explained by semantic refinement alone or GPT-4.1 judge coupling.

arXiv

Calibrating Transformer Attention via Task-Space Sensitivity Feedback

Yawei Liu

Transformer-based pre-trained language models (PLMs) excel in text classification but suffer from attention dilution and attention sink effects, forcing models to over-focus on task-irrelevant tokens. Existing attention supervision methods rely on costly token-level human annotations or static heuristics, which fail to scale or capture context-dependent token importance. To address this, we propose AttCal, a self-supervised, annotation-free attention calibration framework via task-space sensitivity feedback. AttCal treats attention distributions as stochastic policies to guide token deletion perturbations, infers context-dependent token importance from semantic shifts, and optimizes attention parameters via policy gradient. Critically, AttCal freezes the PLM backbone and updates only attention projection layers to ensure training efficiency. Benchmarks across five datasets demonstrate that AttCal significantly outperforms vanilla PLMs-yielding up to a 15.8% accuracy lift for Llama3-8B-and surpasses SOTA supervised baselines in both classification and interpretability.

arXiv

Feedback Loops and Code Perturbations in LLM-based Software Engineering: A Case Study on a C-to-Rust Translation System

Martin Weiss, Jesko Hecking-Harbusch, Jochen Quante, Matthias Woehrle

The advent of strong generative AI has a considerable impact on various software engineering tasks such as code repair, test generation, or language translation. While tools like GitHub Copilot are already in widespread use in interactive settings, automated approaches require a higher level of reliability before being usable in industrial practice. In this paper, we focus on three aspects that directly influence the quality of the results: a) the effect of automated feedback loops, b) the choice of Large Language Model (LLM), and c) the influence of behavior-preserving code changes. We study the effect of these three variables on an automated C-to-Rust translation system. Code translation from C to Rust is an attractive use case in industry due to Rust's safety guarantees. The translation system is based on a generate-and-check pattern, in which Rust code generated by the LLM is automatically checked for compilability and behavioral equivalence with the original C code. For negative checking results, the LLM is re-prompted in a feedback loop to repair its output. These checks also allow us to evaluate and compare the respective success rates of the translation system when varying the three variables. Our results show that without feedback loops LLM selection has a large effect on translation success. However, when the translation system uses feedback loops the differences across models diminish. We observe this for the average performance of the system as well as its robustness under code perturbations. Finally, we also identify that diversity provided by code perturbations can even result in improved system performance.

arXiv

Trust-Based Incentive Mechanisms in Semi-Decentralized Federated Learning Systems

Ajay Kumar Shrestha

In federated learning (FL), decentralized model training allows multi-ple participants to collaboratively improve a shared machine learning model without exchanging raw data. However, ensuring the integrity and reliability of the system is challenging due to the presence of potentially malicious or faulty nodes that can degrade the model's performance. This paper proposes a novel trust-based incentive mechanism designed to evaluate and reward the quality of contributions in FL systems. By dynamically assessing trust scores based on fac-tors such as data quality, model accuracy, consistency, and contribution fre-quency, the system encourages honest participation and penalizes unreliable or malicious behavior. These trust scores form the basis of an incentive mechanism that rewards high-trust nodes with greater participation opportunities and penal-ties for low-trust participants. We further explore the integration of blockchain technology and smart contracts to automate the trust evaluation and incentive distribution processes, ensuring transparency and decentralization. Our proposed theoretical framework aims to create a more robust, fair, and transparent FL eco-system, reducing the risks posed by untrustworthy participants.

arXiv

CUDA-L2: Surpassing cuBLAS Performance for Matrix Multiplication through Reinforcement Learning

Songqiao Su, Xiaoya Li, Albert Wang, Guoyin Wang, Jiwei Li, Chris Shum

In this paper, we propose CUDA-L2, a system that combines large language models (LLMs) and reinforcement learning (RL) to automatically optimize Half-precision General Matrix Multiply (HGEMM) CUDA kernels. Using CUDA execution speed as the RL reward, CUDA-L2 automatically optimizes HGEMM kernels across 1,000 configurations. CUDA-L2 systematically outperforms major matmul baselines to date, from the widely-used torch.matmul to state-of-the-art Nvidia's closed-source libraries, i.e., cuBLAS, cuBLASLt. In offline mode, where kernels are executed consecutively without time intervals, CUDA-L2 yields +22.0% over torch.matmul on average; +19.2% over cuBLAS using the optimal layout configuration (normal-normal NN and transposed-normal TN); +16.8% over cuBLASLt-heuristic, which queries cuBLASLt library and selects the algorithm based on the heuristic's suggestion; and +11.4% over the most competitive cuBLASLt-AutoTuning model, which selects the fastest algorithm from up to 100 candidates from cuBLASLt's suggestions. In server mode, where kernels are executed at random intervals simulating real-time inference, the speedups further increase to +28.7%, +26.0%, +22.4%, and +15.9% for torch.matmul, cuBLAS, cuBLASLt-heuristic, and cuBLASLt-AutoTuning respectively. CUDA-L2 shows that even the most performance-critical, heavily-optimized kernels like HGEMM can be improved through LLM-guided RL automation by systematically exploring configuration spaces at scales impractical for humans. Project and code can be found at github.com/ornith-ai/CUDA-L2

arXiv

Zero-shot reasoning for simulating scholarly peer-review

Khalid M. Saqr

Scholarly publishing requires scalable scrutiny supported by auditable evidence. This paper presents a two-component benchmark of xPeer, the peer-review simulation engine delivered through the xPeerd.com web front-end. The operational component analyzes 352 of 500 simulation records retained under stable-task criteria across disciplines and review modes. The human-reference component releases 1,108 version-1 F1000Research manuscript records with linked human reports and applies a prespecified two-human/two-xPeer comparison. Human-review text and recommendations remained outside the generation input, and source joining occurred after xPeer outputs had been persisted. This procedure defines workflow-level review withholding; prior model exposure falls outside the recorded design. Among 802 records with exactly two human reports, 271 contained two usable xPeer reviewer fields, giving a complete-pair availability rate of 33.8%. Under deterministic extraction rules, median manuscript-level report length was 1,889 words for xPeer and 763 for humans, while median concern count was 41 and 13, respectively. xPeer reports showed higher targeting, category coverage, and executability. Human reports showed higher explicit-reasoning language, lexical manuscript attestation, taxonomy-based scientific relevance, and lower mean within-source redundancy. Cross-source lexical concern matching and recommendation agreement were low. The evidence therefore defines distinct observable review profiles and a transparent reproducibility baseline. Scientific correctness of individual concerns, autonomous editorial use, and cross-system superiority require expert adjudication and common-protocol testing. The study-level dataset and exact version-pinned reproducibility record are archived on Zenodo

arXiv

Trajectory-guided discharge stratification for heart failure using short-context electronic health record sequence modeling

Falk Dippel, Yinan Yu, Annika Rosengren, Martin Lindgren, Christina E. Lundberg, Erik Aerts, Martin Adiels, Helen Sjöland

Purpose: Heart failure (HF) discharge planning depends on identifying patients at risk of deterioration or death, yet accurate prediction from routinely collected electronic health records (EHRs) remains challenging. Methods: We develop trajectory-guided discharge stratification for heart failure (TGDS-HF), a methodology that reads the patient in-hospital trajectory of diagnoses, vital signs, laboratories, medications, and procedures end-to-end with a compact short-context autoregressive Transformer, and uses it to stratify one-year risks of clinical instability (a rehospitalization phenotype) or mortality for discharge care. We instantiate TGDS-HF on a Swedish HF cohort (N = 42,820) to predict one-year clinical instability or mortality at the initial HF diagnosis in-hospital. TGDS-HF has three components: category-level tokenization, recency-weighted temporal representation, and sequence model configuration. We run ablations on these components to show the effectiveness of TGDS-HF. Results: Against traditional eXtreme gradient boosting machine (XGBoost) and bidirectional encoder representations from Transformers (BERT)-based EHR sequence-modeling baselines, TGDS-HF (Llama backbone) achieved area under the precision-recall curves (AUPRCs) with 95% confidence intervals of 0.555 (0.535-0.575) and 0.574 (0.550-0.599) across the two tasks at the method default, with robust calibration. A task-specific refinement using daily aggregation of repeated continuous events improves the mortality task to 0.582 (0.558-0.608). Further, TGDS-HF maintains strong performance under reduced clinical concept availability and limited training data. Conclusion: Combined predictions of instability and mortality from TGDS-HF may support personalized discharge planning, ranging from follow-up in primary care to specialist-led management and, when appropriate, palliative care.

arXiv

MermaidSeqBench: An Evaluation Benchmark for NL-to-Mermaid Sequence Diagram Generation

Basel Shbita, Farhan Ahmed, Chad DeLuca

Large language models (LLMs) have shown great promise in generating structured diagrams from natural language descriptions, particularly Mermaid sequence diagrams for software engineering. However, the lack of existing benchmarks to assess the LLM's correctness on this task hinders rigorous, systematic evaluation and principled comparison of model capabilities on this task. To address this shortcoming, we introduce MermaidSeqBench, a human-verified and synthetically extended benchmark for assessing LLM capabilities in generating Mermaid sequence diagrams from natural language prompts. The benchmark consists of 132 samples developed via a hybrid methodology of human-verified flows, LLM-based augmentation, and rule-based expansion. The evaluation uses an LLM-as-a-judge model to assess generation across various fine-grained metrics such as syntax correctness, activation handling, error handling, and practical usability. To demonstrate the effectiveness and flexibility of our benchmark, we perform initial evaluations on numerous state-of-the-art LLMs with multiple LLM judges which reveal significant capability gaps across models and evaluation modes. MermaidSeqBench provides a foundation for evaluating structured diagram generation and advances the scientific understanding of LLM capabilities and limitations in structured generation tasks.

arXiv

Worst-Case Distance-Aware Error Bounds for Neural Networks

Masoud Ataei, Vikas Dhiman, Mohammad Javad Khojasteh

Safety-critical applications of machine learning require uncertainty estimates that support reliable worst-case analysis. Neural networks (NNs) provide expressive function approximation, while Gaussian processes (GPs) offer principled probabilistic uncertainty, but both face limitations in this setting: most modern neural architectures lack tractable worst-case error bounds, and Gaussian processes become computationally expensive at scale. For uncertainty to be interpretable, a central requirement is distance-awareness: that is, uncertainty increases with the distance between a test input and the nearest relevant training data. We present a general framework for worst-case distance-aware error bounds for NNs that combine dense layers with spline-based components. Our approach establishes error bounds that are both distance-aware, reflecting proximity of a test point to its nearest training data, and worst-case, providing deterministic guarantees under known Lipschitz constraints rather than probabilistic assumptions. Our algorithm, K-DAREK (Distance-Aware Error for Kurkova-Kolmogorov Networks), provides efficient and interpretable uncertainty quantification for NNs. K-DAREK is about four times faster and ten times more computationally efficient than an ensemble of KANs, 8.6 times more scalable than GP, eliminates up to 8.2% error-bound violation rate observed in DAREK, and reduces the average collision rate from 1.8% to 1.1% in the multi-agent safe control experiment. On high-dimensional real-world regression tasks (e.g., Real Estate Valuation), K-DAREK preserves distance-aware error bounds and achieves zero coverage violations, addressing the overgeneralization and inducing-point-coverage limitations exhibited by SNGP and DUE, respectively.

arXiv

GRALS: GCN-Guided Redundancy-Aware Local Search for Minimum Vertex Cover

Chanjuan Liu, Qiqi Bao, Yu Zhang, Enqiang Zhu

The minimum vertex cover (MVC) problem seeks to identify the smallest set of vertices that cover all edges in an undirected graph. As a fundamental NP-hard combinatorial optimization problem, MVC has been widely studied due to its applications in network analysis and system design. For large-scale instances, local search heuristics are among the most effective approaches, offering a practical trade-off between solution quality and computational efficiency. Most existing high-performance local search algorithms adopt a break-and-repair framework, where the repair process mainly restores feasibility but provides limited opportunities for enhancing the current solution's structure. We introduce GRALS, a local search framework that integrates vertex probability priors learned from a graph convolutional network with an expansion revelation elimination (ERE) operator. The learned priors guide the search toward promising regions, while the ERE operator dynamically identifies and eliminates redundant vertices to further enhance solution quality. Experiments conducted on 346 benchmark instances demonstrate that GRALS achieves the best-known solution for 335 instances, compared to 302 instances achieved by the strongest baseline. The advantage of GRALS is maintained across different time limits and is particularly evident on large-scale graphs with millions of vertices.

arXiv

Autonomous Learning From Success and Failure: Goal-Conditioned Supervised Learning with Negative Feedback

Zeqiang Zhang, Fabian Wurzberger, Gerrit Schmid, Sebastian Gottwald, Daniel A. Braun

Learning from reward functions and imitation learning of demonstrations are the two principal approaches for training autonomous systems that interact with an environment through action and observation. Both, however, require human specification for each behaviour to be acquired, a problem for long-lived self-adaptive systems whose goals and operating conditions cannot be fully anticipated at design time. Recently, Goal-Conditioned Supervised Learning (GCSL) through self-imitation has been proposed as a self-supervised alternative: by strategically relabelling goals, agents can derive policy insights from their own experiences. Despite its successes, this framework presents two notable limitations: (1) learning exclusively from self-generated experiences can exacerbate the agents' inherent biases; (2) the relabelling strategy allows agents to focus solely on successful outcomes, precluding them from learning from their mistakes. To address these issues, we propose GCSL with Negative Feedback (GCSL-NF), which evaluates each trajectory twice: positively with respect to relabelled goals, and correctively with respect to the goal originally intended. The corrective target comes from a similarity function learned contrastively from trajectory-induced neighbourhood relations, so that neither a reward function nor a geometric distance needs to be specified. Our experiments show that GCSL-NF overcomes limitations imposed by agents' initial biases, increasingly benefits from negative feedback as learning progresses, and matches or surpasses GCSL- and HER-based methods. By reducing reliance on prespecified reward functions, the proposed approach is particularly relevant for self-adaptive autonomous systems, where adaptation objectives may be diverse, changing, or difficult to engineer.

arXiv

On The Suitability of Differential Dataflow For Datalog Interpretation In Highly Dynamic Settings

Bruno Rucy Carneiro Alves de Lima, Merlin Kramer, Kalmer Apinis

In the domain of knowledge representation and reasoning within AI, datalog engines play an ever-increasingly crucial role. The crux of their operation lies in materialization: the evaluation of a data- log program and its incorporation into a database. This operation becomes complex and resource-intensive, especially when the data is highly dynamic, as it is common in distributed environments. Thus, incremental materialization, adjusting the computation to new data instead of restarting it, is the norm. However, handling the deletion of data is significantly more complicated than addition due to the cascading effects of what is being removed. Differential Dataflow offers a computational model that effectively addresses this, ensuring consistent performance for both data additions and deletions. In this paper, we delve into the efficiency of materializa- tion using three distinct datalog implementations: one based on a streamlined relational engine and two others that implement the same algorithm, but with one utilizing differential-dataflow, and another not. Our insights provide a roadmap for enhancing datalog- driven computations, particularly in dynamic data environments like the cloud.

arXiv

A Reverse-BSDE Diffusion Sampler

Jairon H. N. Batista, Flávio B. Gonçalves, Yuri F. Saporito, Rodrigo S. Targino

Diffusion-based generative models have renewed interest in stochastic differential equation methods for sampling from complex distributions. We study a setting in which the target density is known only up to a normalizing constant and reformulate the reverse-time diffusion sampler as a forward-backward stochastic differential equation (FBSDE). This formulation replaces the separate pre-estimation of the time-dependent score with the solution of a coupled stochastic system. We prove the equivalence of these formulations and provide a decomposition of the approximation error arising from initializing the sampler with a standard Gaussian, applying Euler discretization to the dynamics, and solving the FBSDE approximately. We then evaluate the proposed algorithm on synthetic targets, including separated mixtures, anisotropic Gaussian distributions, and banana-shaped and ring-shaped distributions. The results demonstrate the promise of the method, particularly for targets with complex global structure.

arXiv

ASAT: Adaptive Scoring and Thresholding with Human Feedback for Robust Out-of-Distribution Detection

Daisuke Yamada, Harit Vishwakarma, Ramya Korlakai Vinayak

Machine Learning (ML) models are trained on in-distribution (ID) data but often encounter out-of-distribution (OOD) inputs during deployment---posing serious risks in safety-critical domains. Recent works have focused on designing scoring functions to quantify OOD uncertainty, with score thresholds typically set based solely on ID data to achieve a target true positive rate (TPR), since OOD data is limited before deployment. However, these TPR-based thresholds leave false positive rates (FPR) uncontrolled, often resulting in high FPRs where OOD points are misclassified as ID. Moreover, fixed scoring functions and thresholds lack the adaptivity needed to handle newly observed, evolving OOD inputs, leading to sub-optimal performance. To address these challenges, we propose \emph{ASAT}, a human-in-the-loop framework that \textit{safely updates both scoring functions and thresholds on the fly} based on real-world OOD inputs. ASAT maximizes TPR while controlling FPR at all times under stationary conditions, even as the system adapts over time. Under nonstationary conditions, the method adapts to distribution shifts with only transient FPR violations during the adaptation period. We provide theoretical guarantees for FPR control under stationary conditions and present extensive empirical evaluations on OpenOOD benchmarks to demonstrate that our approach outperforms existing methods by achieving higher TPRs while maintaining FPR control.

arXiv

OctoLong: Mid-Training On Cross-Repository Code Contexts Enhances Long-Context Modeling

Indraneil Paul, Falko Helm, Goran Glavaš, Iryna Gurevych

Context lengths of language models (LMs) have dramatically increased, driven by the demands for in-context learning, self-improvement, and long-horizon agentic workflows. Existing long-context corpora, however, are dominated by books, academic articles, and code repositories, which are finite resources and often scarce in long-distance dependencies. In this work, we introduce OctoLong, a context engineering pipeline that instruments an AST parser, a language server backend, and a package manager to facilitate the recursive retrieval of code references, enabling the curation of dependency-rich code contexts of millions of tokens in length. We then train OctoLong-Instruct, a suite of capable long-context open LMs, derived from base models ranging in size from 600M to 14B parameters, via context-extension mid-training on a ~50B-token mixture containing ~6.2B tokens of OctoLong code contexts, followed by ~10B tokens of instruction tuning. Our training ablations and experimental evaluations against 18 state-of-the-art open-weight long-context LMs show that supplanting just 12% of traditional context-extension corpora with OctoLong data yields substantial gains in long-range retrieval, long-term state tracking, repository-level code understanding, and downstream agentic tasks, while also enhancing API usage in short-context coding scenarios.

arXiv

DASyR-LLM: Domain-Aware Symbolic Regression with LLMs for Kinetic Model Discovery

Roberto Aliaga Medina, Paulina Quintanilla, Antonio Del rio chanona

Kinetic model discovery is a central challenge in chemical engineering, as accurate rate expressions are essential for understanding and controlling chemical and biological processes. Symbolic regression (SR) has emerged as a powerful data-driven approach for identifying interpretable kinetic models, but usually operates without domain knowledge, often exploring physicochemically implausible models. Large language models (LLMs) offer a promising avenue for injecting domain expertise into this search. Here, we introduce an LLM-guided SR framework, embedding an LLM module within an iterative SR algorithm for automated kinetic model discovery. The LLM performs two roles at each iteration: (1) a qualitative physicochemical critique of the best SR candidates, and (2) the proposal of new candidate rate expressions guided by the SR-generated models and embedded chemical knowledge. Our framework is evaluated on four in silico case studies of increasing complexity, spanning heterogeneous catalysis and bioprocess systems. Results show the LLM-guided framework reduces iterations to identify the ground-truth model by $41.7-79.3\%$ versus a state-of-the-art SR framework, with the LLM directly proposing the correct model structure in over half of the guided runs. In practical settings, where each iteration typically requires a new wet-lab experiment, this translates into a substantial reduction in experimental effort. Predictive performance on an independent validation set is equivalent between both approaches, with $R^2>0.98$ in all case studies. Ablation studies indicate that both the SR component and the LLM scale contribute to this performance, with a reduced-size LLM largely retaining discovery efficiency. These findings demonstrate that LLMs can effectively inject domain knowledge into scientific model discovery, paving the way toward fully automated, domain-aware kinetic modelling pipelines.

arXiv

Stable Density Ridges: Consistency and Convergence of Subspace Constrained Mean Shift

Wanli Qiao

The Subspace Constrained Mean Shift (SCMS) algorithm is a popular nonparametric method for extracting density ridges, which serve as a low-dimensional representation of high-dimensional data. It is a widely held belief in the literature that SCMS trajectories converge to the classical density ridge, which we call the "static ridge", defined via the density gradient and the eigenvalues and eigenvectors of the density's Hessian. In this paper, we demonstrate that this assumption does not hold in general, as the static definition fails to account for the rotation of the trailing eigenspace along the continuous flow of the algorithm's underlying vector field. To resolve this, we propose a paradigm shift by introducing the "stable ridge", a novel geometric structure defined through the lens of dynamical systems and the Jacobian of the projected density gradient. We prove that this stable ridge is the true theoretical target of the SCMS algorithm. Building upon this foundation, we develop a generalized SCMS framework utilizing a constant step size, establishing its uniform R-linear convergence and topological surjectivity onto the stable ridge. We further derive the rates of convergence for estimating the stable ridge in terms of the Hausdorff distance. Finally, we expose that the original SCMS algorithm suffers from polynomial-time computational complexity, which is caused by implicitly coupling the step size to the smoothing bandwidth via the Mean Shift operator, and demonstrate how our generalized framework provides a statistically consistent and more efficient solution.

arXiv

Reward Structure Shapes the Interaction Between Episodic Exploration and Neural Memory in Reinforcement Learning

Jai Malegaonkar, Rohan Patil, Henrik I. Christensen

In partially observable reinforcement learning, agents face a dual bottleneck: they must explore to encounter rewarding states and retain that experience in memory to optimize their policies. Exploration bonuses and memory architectures are traditionally evaluated in isolation, leaving their interaction unmeasured, and standard notions of sparse reward conflate temporal signal density with what the reward actually supervises. We present a controlled study crossing episodic exploration bonuses with diverse neural memory architectures across three environments that vary how the content of memory is acquired. An identical bonus signal yields three distinct interaction patterns: it amplifies architectural capacity differences where memory content must be actively discovered and retained unsupervised; equalizes architectures to a shared ceiling where the content, once sought out, is a single reward-supervised cue; and is null where the observation stream is purely scheduled. Controlled reward manipulations verify that these patterns track reward structure rather than density: a dense reward neutralizes a bonus only if it directly supervises the required latent memory, and a small avoidable penalty on exploratory actions (leaving the optimum unchanged) induces policy convergence to suboptimal stationary states, which either bonus resolves. We then formalize reward sparsity with observation-anchored reward machines, separating structural sparsity (an automaton reproduces the return without the task-required history) from potential sparsity (the one-step reward misprices local exploratory actions); the resulting vocabulary organizes the three regimes by the retention burden each task exposes. Together, these results show exploration and memory are complements, not substitutes: a bonus induces exposure, and only memory converts exposure into return.

arXiv

Representational separation between unitary and channel quantum generative models via shared classical randomness at shallow depth

Arunava Majumder, Marius Krumm, Hendrik Poulsen Nautrup, Hans J Briegel

Near-term quantum hardware limits circuit depth and often imposes geometrically local connectivity for quantum generative models, restricting the output distributions accessible to shallow unitary Born models. Introducing stochasticity into a unitary quantum Born model can improve the empirical generative performance of the resulting channel model and, for a restricted small-scale architecture, has been proven to represent a strictly larger family of distributions than its unitary counterpart. However, whether such randomness provides a provable separation at fixed shallow depth for arbitrarily large systems has remained open. Here, we show that shared classical randomness, a comparatively weak resource from entanglement theory, is sufficient to establish such a strict scalable representational separation over the corresponding shallow unitary Born model. More specifically, we augment bounded-connectivity shallow unitary circuits, followed by computational-basis measurements, with spatially separated local Pauli operations, whose joint application is controlled by a single classically sampled random bit. The resulting shallow-depth channel model generates long-range correlations in the classical output distribution that no purely unitary shallow-depth model with bounded connectivity can reproduce. For one-dimensional nearest-neighbour architectures, reproducing such distributions with a purely unitary model can require depth $Ω(N)$ in the worst case. We further show that measurement-based quantum computation (MBQC) provides a natural implementation of the required shared classical randomness through suitable adaptation of the random measurement outcomes. Numerical experiments on MBQC-based generative models support the analytical results.

arXiv

BnBERT-iPET: Sparse Few-Shot Language Modeling for Bengali via Lottery Ticket Pruning

Sajib Hossain, Md Kamrus Samad, Anan Ghosh, Labib Imam Chowdhury, Nabeel Mohammed

Deep neural networks have shown impressive success in NLP tasks owing to their complex structure and huge number of edges. Achieving state-of-the-art performance in natural language processing with a large pre-trained model such as BERT is expensive and time-consuming, carries a large carbon footprint, and is difficult to realize on machines with minimal computational capability. This creates a barrier to training complex models for resource-constrained languages such as Bengali. However, in a complex neural model, not all edges are equally impactful, and the contributions of some of them can be neglected. Pruning promises to reduce the memory footprint of regular networks, shorten the training time of ever-growing networks, and increase inference efficiency without sacrificing comparable performance. In this work, we introduce BnBERT-iPET, a sparse few-shot language modeling approach for Bengali, and experimentally show that a lightweight few-shot-learned language model retaining only 10% of the edges of an initial model such as BERT can perform neck and neck with much larger models on challenging tasks for a resource-constrained language such as Bengali. By learning from few shots through iterative pattern exploiting training and achieving 90% sparsity with the Lottery Ticket Hypothesis pruning technique, our pruned BnBERT-iPET model proves to be a tough competitor to state-of-the-art language models such as Bangla Electra, Indic-BERT, and XLM-RoBERTa on downstream tasks over standard benchmark datasets of the Bengali language.

Keep browsing

More papers from top AI conferences

30+ top conferences, by conference, year, and topic.