Given a single observation from a Gaussian distribution with unknown mean $θ$, we design computationally efficient procedures that can approximately generate an observation from a different target distribution $Q_θ$ uniformly for all $θ$ in a parameter set. We leverage our technique to establish reduction-based computational lower bounds for several canonical high-dimensional statistical models under widely-believed conjectures in average-case complexity. In particular, we cover cases in which: 1. $Q_θ$ is a general location model with non-Gaussian distribution, including both light-tailed examples (e.g., generalized normal distributions) and heavy-tailed ones (e.g., Student's $t$-distributions). As a consequence, we show that computational lower bounds proved for spiked tensor PCA with Gaussian noise are universal, in that they extend to other non-Gaussian noise distributions within our class. 2. $Q_θ$ is a normal distribution with mean $f(θ)$ for a general, smooth, and nonlinear link function $f:\mathbb{R} \rightarrow \mathbb{R}$. Using this reduction, we construct a reduction from symmetric mixtures of linear regressions to generalized linear models with link function $f$, and establish computational lower bounds for solving the $k$-sparse generalized linear model when $f$ is an even function. This result constitutes the first reduction-based confirmation of a $k$-to-$k^2$ statistical-to-computational gap in $k$-sparse phase retrieval, resolving a conjecture posed by Cai et al. (2016). As a second application, we construct a reduction from the sparse rank-1 submatrix model to the planted submatrix model, establishing a pointwise correspondence between the phase diagrams of the two models that faithfully preserves regions of computational hardness and tractability.

In supervised classification tasks, models are trained to predict a label for each data point. In real-world datasets, these labels are often noisy due to annotation errors. While the impact of label noise on the performance of deep learning models has been widely studied, its effects on the networks' hidden representations remain poorly understood. We address this gap by systematically comparing hidden representations using the Information Imbalance, a computationally efficient proxy of conditional mutual information. Through this analysis, we observe that the information content of the hidden representations follows a double descent as a function of the number of network parameters, akin to the behavior of the test error. We further demonstrate that in the underparameterized regime, representations learned with noisy labels are more informative than those learned with clean labels, while in the overparameterized regime, these representations are equally informative. Our results indicate that the representations of overparameterized networks are robust to label noise. We also found that the information imbalance between the penultimate and pre-softmax layers decreases with cross-entropy loss in the overparameterized regime. This offers a new perspective on understanding generalization in classification tasks. Extending our analysis to representations learned from random labels, we show that these perform worse than random features. This indicates that training on random labels drives networks much beyond lazy learning, as weights adapt to encode labels information.

Abstract--The 2024 July Revolution in Bangladesh represents a landmark event in the study of civil resistance: a successful, student-led civilian uprising that overthrew a long-standing authoritarian regime despite facing brutal state repression. This study investigates the central paradox of its success: how state violence, intended to quell dissent, ultimately fueled the movement's victory. We employ a mixed-methods approach. First, we develop a qualitative narrative of the conflict's timeline to generate specific, testable hypotheses. Then, using a disaggregated, event-level dataset, we employ a multi-method quantitative analysis to dissect the complex relationship between repression and mobilisation. We provide a framework to analyse explosive modern uprisings like the July Revolution. Initial pooled regression models highlight the crucial role of protest momentum (measured by a feedback loop effect) in sustaining the movement. T o isolate causal effects, we specify a Two-Way Fixed Effects panel model, which provides robust evidence for a direct and statistically significant local suppression backfire effect. Our V ector Autoregression (V AR) analysis provides clear visual evidence of an immediate, nationwide mobilisation in response to increased lethal violence. We further demonstrate that this effect was non-linear . A structural break analysis reveals that the backfire dynamic was statistically insignificant in the conflict's early phase but was triggered by the catalytic moral shock of the first wave of lethal violence, and its visuals circulated around July 16th. We conclude that the July Revolution was driven by a contingent, non-linear backfire, triggered by specific catalytic moral shocks and accelerated by the viral reaction to the visual spectacle of state brutality. N August 2024, the fifteen-year rule of Prime Minister Sheikh Hasina of Bangladesh came to a sudden and dramatic end. After weeks of escalating nationwide protests, she resigned from her post and fled the country. These authors contributed equally to this work. Saiful Bari Siddiqui is a Senior Lecturer at the Department of Computer Science and Engineering, BRAC University, Dhaka, Bangladesh (e-mail: saiful.bari@bracu.ac.bd). Anupam Debashis Roy is a PhD candidate at the Department of Sociology, University of Oxford, Oxford, United Kingdom (e-mail: anu-pam.roy@sant.ox.ac.uk). In a matter of weeks, this initial spark grew into a nationwide fire, as hundreds of thousands of ordinary citizens joined the students, bringing the country to a standstill and achieving a political transformation that had seemed unthinkable just a month earlier.

The development of Automatic Speech Recognition (ASR) systems for low-resource African languages remains challenging due to limited transcribed speech data. While recent advances in large multilingual models like OpenAI's Whisper offer promising pathways for low-resource ASR development, critical questions persist regarding practical deployment requirements. This paper addresses two fundamental concerns for practitioners: determining the minimum data volumes needed for viable performance and characterizing the primary failure modes that emerge in production systems. We evaluate Whisper's performance through comprehensive experiments on two Bantu languages: systematic data scaling analysis on Kinyarwanda using training sets from 1 to 1,400 hours, and detailed error characterization on Kikuyu using 270 hours of training data. Our scaling experiments demonstrate that practical ASR performance (WER < 13\%) becomes achievable with as little as 50 hours of training data, with substantial improvements continuing through 200 hours (WER < 10\%). Complementing these volume-focused findings, our error analysis reveals that data quality issues, particularly noisy ground truth transcriptions, account for 38.6\% of high-error cases, indicating that careful data curation is as critical as data volume for robust system performance. These results provide actionable benchmarks and deployment guidance for teams developing ASR systems across similar low-resource language contexts. We release accompanying and models see https://github.com/SunbirdAI/kinyarwanda-whisper-eval

There are more than 2000 living languages in Africa, most of which have been bypassed by advances in language technology. Current leading LLMs exhibit strong performance on a number of the most common languages (e.g. Swahili or Yoruba), but prioritise support for the languages with the most speakers first, resulting in piecemeal ability across disparate languages. We contend that a regionally focussed approach is more efficient, and present a case study for Uganda, a country with high linguistic diversity. We describe the development of Sunflower 14B and 32B, a pair of models based on Qwen 3 with state of the art comprehension in the majority of all Ugandan languages. These models are open source and can be used to reduce language barriers in a number of important practical applications.

Anomaly detection is a key task across domains such as industry, healthcare, and cybersecurity. Many real-world anomaly detection problems involve analyzing multiple features over time, making time series analysis a natural approach for such problems. While deep learning models have achieved strong performance in this field, their trend to exhibit high energy consumption limits their deployment in resource-constrained environments such as IoT devices, edge computing platforms, and wearables. To address this challenge, this paper introduces the \textit{Vacuum Spiker algorithm}, a novel Spiking Neural Network-based method for anomaly detection in time series. It incorporates a new detection criterion that relies on global changes in neural activity rather than reconstruction or prediction error. It is trained using Spike Time-Dependent Plasticity in a novel way, intended to induce changes in neural activity when anomalies occur. A new efficient encoding scheme is also proposed, which discretizes the input space into non-overlapping intervals, assigning each to a single neuron. This strategy encodes information with a single spike per time step, improving energy efficiency compared to conventional encoding methods. Experimental results on publicly available datasets show that the proposed algorithm achieves competitive performance while significantly reducing energy consumption, compared to a wide set of deep learning and machine learning baselines. Furthermore, its practical utility is validated in a real-world case study, where the model successfully identifies power curtailment events in a solar inverter. These results highlight its potential for sustainable and efficient anomaly detection.

We study reinforcement learning (RL) fine-tuning of large language model (LLM) agents for long-horizon multi-turn tool use, where context length quickly becomes a fundamental bottleneck. Existing RL pipelines can suffer from degraded instruction following, excessive rollout costs, and most importantly, strict context limits. To address these challenges, we introduce summarization-based context management to training. In specific, it periodically compresses the tool using history by LLM-generated summaries that retain task-relevant information to keep a compact context while enabling the agent to scale beyond the fixed context window. Building on this formulation, we derive a policy gradient representation that seamlessly enables standard LLM RL infrastructures to optimize both tool-use behaviors as well as summarization strategies in an end-to-end fashion. We instantiate this framework with \underline{SU}mmarization augmented \underline{P}olicy \underline{O}ptimization (\texttt{SUPO}), an LLM RL algorithm that enables long-horizon training beyond a fixed context limit. Experiments on interactive function calling and searching tasks demonstrate that \texttt{SUPO} significantly improves the success rate while maintaining the same or even lower working context length compared to baselines. We also demonstrate that for complex searching tasks, \texttt{SUPO} can further improve the evaluation performance when scaling test-time maximum round of summarization beyond that of training time. Our results establish summarization-based context management as a principled and scalable approach for training RL agents beyond a fixed context length limit.

Reward models (RMs) are central to the alignment of language models (LMs). An RM often serves as a proxy for human preferences to guide downstream LM behavior. However, our understanding of RM behavior is limited. Our work (i) formalizes a framework for measuring the alignment of opinions captured by RMs, (ii) investigates the extent to which RMs demonstrate sociodemographic biases, and (iii) explores the effects of prompting to steer rewards towards the preferences of a target group. We study the subjective and diverse perspectives on controversial topics, which allows us to quantify RM perspectives in terms of their opinions, attitudes, and values. We show that RMs are poorly aligned with several demographic groups and can systematically reward harmful stereotypes, and steering alone is not enough to overcome these limitations. Our findings underscore the need for more careful consideration of RM behavior in model alignment during preference learning to prevent the propagation of unwanted social biases in the language technologies that we use.

Large-scale multimodal models achieve strong results on tasks like Visual Question Answering (VQA), but they often fail when queries require culturally grounded, everyday knowledge, particularly in low-resource and underrepresented languages. QA ( EverydayMMQA), a framework for creating large-scale, culturally-grounded datasets for spoken and visual question answering (SVQA). Using this framework, we developed OASIS, a multimodal dataset integrating speech, images, and text. With over 0.92M images and 14.8M QA pairs, OASIS contains 3.7M spoken questions, enabling four unique input combinations: speech-only, text-only, speech+image, and text+image. Focused on English and Arabic varieties, 18 countries, the dataset content is curated to reflect diverse, real-world situations. OASIS tests models on tasks beyond object recognition that involve pragmatic, commonsense, and culturally aware reasoning. EverydayMMQA and OASIS together provide a benchmark and training dataset for building multimodal LLMs for a comprehensive set of everyday tasks within cultural contexts. The framework and dataset will be made publicly available to the community. This multi-sensory integration is fundamental to how humans understand the surroundings and communicate. As large language models (LLMs) evolve, it is important to train them with multiple modalities: speech, text, and images, to mimic human interaction. For instance, when asking about an object, we often point to it while asking a question. In this scenario, we expect an AI assistant to process a multimodal triplet: the visual information (what we're pointing at), the spoken information (our question), and the contextual knowledge required to provide a culturally appropriate response (see Figure 1). Crucially, this contextual knowledge is not universal: it is shaped by culture and language.

We present EvalMORAAL, a transparent chain-of-thought (CoT) framework that uses two scoring methods (log-probabilities and direct ratings) plus a model-as-judge peer review to evaluate moral alignment in 20 large language models. We assess models on the World Values Survey (55 countries, 19 topics) and the PEW Global Attitudes Survey (39 countries, 8 topics). With EvalMORAAL, top models align closely with survey responses (Pearson's r approximately 0.90 on WVS). Yet we find a clear regional difference: Western regions average r=0.82 while non-Western regions average r=0.61 (a 0.21 absolute gap), indicating consistent regional bias. Our framework adds three parts: (1) two scoring methods for all models to enable fair comparison, (2) a structured chain-of-thought protocol with self-consistency checks, and (3) a model-as-judge peer review that flags 348 conflicts using a data-driven threshold. Peer agreement relates to survey alignment (WVS r=0.74, PEW r=0.39, both p<.001), supporting automated quality checks. These results show real progress toward culture-aware AI while highlighting open challenges for use across regions.