Do Taliban's drone attacks expose a chink in Pakistan's armour? On the evening of March 13, drones struck three locations across Pakistan. Two children were wounded in Quetta. Civilians were also injured in Kohat and in Rawalpindi, the garrison city that houses the headquarters of Pakistan's armed forces and neighbours the capital, Islamabad. Pakistan's military said the drones were intercepted before reaching their targets.

This study investigates the impact of localized updates to large language models (LLMs), specifically in the context of knowledge editing - a task aimed at incorporating or modifying specific facts without altering broader model capabilities. We first show that across different post-training interventions like continuous pre-training, full fine-tuning and LORA-based fine-tuning, the Frobenius norm of the updated matrices always increases. This increasing norm is especially detrimental for localized knowledge editing, where only a subset of matrices are updated in a model . We reveal a consistent phenomenon across various editing techniques, including fine-tuning, hypernetwork-based approaches, and locate-and-edit methods: the norm of the updated matrix invariably increases with successive updates. Such growth disrupts model balance, particularly when isolated matrices are updated while the rest of the model remains static, leading to potential instability and degradation of downstream performance. Upon deeper investigations of the intermediate activation vectors, we find that the norm of internal activations decreases and is accompanied by shifts in the subspaces occupied by these activations, which shows that these activation vectors now occupy completely different regions in the representation space compared to the unedited model. With our paper, we highlight the technical challenges with continuous and localized sequential knowledge editing and their implications for maintaining model stability and utility.

Automatically summarizing large text collections is a valuable tool for document research, with applications in journalism, academic research, legal work, and many other fields. In this work, we contrast two classes of systems for large-scale multi-document summarization (MDS): compression and full-text. Compression-based methods use a multi-stage pipeline and often lead to lossy summaries. Full-text methods promise a lossless summary by relying on recent advances in long-context reasoning. To understand their utility on large-scale MDS, we evaluated them on three datasets, each containing approximately one hundred documents per summary. Our experiments cover a diverse set of long-context transformers (Llama-3.1, Command-R, Jamba-1.5-Mini) and compression methods (retrieval-augmented, hierarchical, incremental). Overall, we find that full-text and retrieval methods perform the best in most settings. With further analysis into the salient information retention patterns, we show that compression-based methods show strong promise at intermediate stages, even outperforming full-context. However, they suffer information loss due to their multi-stage pipeline and lack of global context. Our results highlight the need to develop hybrid approaches that combine compression and full-text approaches for optimal performance on large-scale multi-document summarization.

As events progress, news articles often update with new information: if we are not cautious, we risk propagating outdated facts. In this work, we hypothesize that linguistic features indicate factual fluidity, and that we can predict which facts in a news article will update using solely the text of a news article (i.e. not external resources like search engines). We test this hypothesis, first, by isolating fact-updates in large news revisions corpora. News articles may update for many reasons (e.g. factual, stylistic, narrative). We introduce the NewsEdits 2.0 taxonomy, an edit-intentions schema that separates fact updates from stylistic and narrative updates in news writing. We annotate over 9,200 pairs of sentence revisions and train high-scoring ensemble models to apply this schema. Then, taking a large dataset of silver-labeled pairs, we show that we can predict when facts will update in older article drafts with high precision. Finally, to demonstrate the usefulness of these findings, we construct a language model question asking (LLM-QA) abstention task. We wish the LLM to abstain from answering questions when information is likely to become outdated. Using our predictions, we show, LLM absention reaches near oracle levels of accuracy.

We present LEGAL-UQA, the first Urdu legal question-answering dataset derived from Pakistan's constitution. This parallel English-Urdu dataset includes 619 question-answer pairs, each with corresponding legal article contexts, addressing the need for domain-specific NLP resources in low-resource languages. We describe the dataset creation process, including OCR extraction, manual refinement, and GPT-4-assisted translation and generation of QA pairs. Our experiments evaluate the latest generalist language and embedding models on LEGAL-UQA, with Claude-3.5-Sonnet achieving 99.19% human-evaluated accuracy. We fine-tune mt5-large-UQA-1.0, highlighting the challenges of adapting multilingual models to specialized domains. Additionally, we assess retrieval performance, finding OpenAI's text-embedding-3-large outperforms Mistral's mistral-embed. LEGAL-UQA bridges the gap between global NLP advancements and localized applications, particularly in constitutional law, and lays the foundation for improved legal information access in Pakistan.

We consider solving complex spatiotemporal dynamical systems governed by partial differential equations (PDEs) using frequency domain-based discrete learning approaches, such as Fourier neural operators. Despite their widespread use for approximating nonlinear PDEs, the majority of these methods neglect fundamental physical laws and lack interpretability. We address these shortcomings by introducing Physics-embedded Fourier Neural Networks (PeFNN) with flexible and explainable error control. PeFNN is designed to enforce momentum conservation and yields interpretable nonlinear expressions by utilizing unique multi-scale momentum-conserving Fourier (MC-Fourier) layers and an element-wise product operation. The MC-Fourier layer is by design translation- and rotation-invariant in the frequency domain, serving as a plug-and-play module that adheres to the laws of momentum conservation. PeFNN establishes a new state-of-the-art in solving widely employed spatiotemporal PDEs and generalizes well across input resolutions. Further, we demonstrate its outstanding performance for challenging real-world applications such as large-scale flood simulations.

Codebooks -- documents that operationalize constructs and outline annotation procedures -- are used almost universally by social scientists when coding unstructured political texts. Recently, to reduce manual annotation costs, political scientists have looked to generative large language models (LLMs) to label and analyze text data. However, previous work using LLMs for classification has implicitly relied on the universal label assumption -- correct classification of documents is possible using only a class label or minimal definition and the information that the LLM inductively learns during its pre-training. In contrast, we argue that political scientists who care about valid measurement should instead make a codebook-construct label assumption -- an LLM should follow the definition and exclusion criteria of a construct/label provided in a codebook. In this work, we collect and curate three political science datasets and their original codebooks and conduct a set of experiments to understand whether LLMs comply with codebook instructions, whether rewriting codebooks improves performance, and whether instruction-tuning LLMs on codebook-document-label tuples improves performance over zero-shot classification. Using Mistral 7B Instruct as our LLM, we find re-structuring the original codebooks gives modest gains in zero-shot performance but the model still struggles to comply with the constraints of the codebooks. Optimistically, instruction-tuning Mistral on one of our datasets gives significant gains over zero-shot inference (0.76 versus 0.53 micro F1). We hope our conceptualization of the codebook-specific task, assumptions, and instruction-tuning pipeline as well our semi-structured LLM codebook format will help political scientists readily adapt to the LLM era.

This paper critically discusses how generative artificial intelligence (GenAI) might impose Western ideologies on non-Western societies, perpetuating digital neocolonialism in education through its inherent biases. It further suggests strategies for local and global stakeholders to mitigate these effects. Our discussions demonstrated that GenAI can foster cultural imperialism by generating content that primarily incorporates cultural references and examples relevant to Western students, thereby alienating students from non-Western backgrounds. Also, the predominant use of Western languages by GenAI can marginalize non-dominant languages, making educational content less accessible to speakers of indigenous languages and potentially impacting their ability to learn in their first language. Additionally, GenAI often generates content and curricula that reflect the perspectives of technologically dominant countries, overshadowing marginalized indigenous knowledge and practices. Moreover, the cost of access to GenAI intensifies educational inequality and the control of GenAI data could lead to commercial exploitation without benefiting local students and their communities. We propose human-centric reforms to prioritize cultural diversity and equity in GenAI development; a liberatory design to empower educators and students to identify and dismantle the oppressive structures within GenAI applications; foresight by design to create an adjustable GenAI system to meet future educational needs; and finally, effective prompting skills to reduce the retrieval of neocolonial outputs.

Iran and Pakistan have carried out air attacks on each other's territories, targeting armed groups near their 900km-long (559-mile) volatile border, which they say were meant to ensure their respective national security. Iran's powerful Islamic Revolutionary Guard Corps (IRGC) targeted an armed group in Panjgur town of Pakistan's Balochistan province late on Tuesday, prompting Pakistan to bomb hideouts of armed groups in the Sistan-Baluchestan province of Iran early on Thursday. Let's take a look at why the neighbours have resorted to direct military strikes, who the targets were, and what the attacks tell us. The IRGC, an elite force which is a vital part of the Iranian establishment but separate from Iran's army, hit the Jaish al-Adl armed group with missile and drone strikes in a mountainous region in Pakistan close to the Iranian border. Iran said it targeted the Iranian "terrorist" group it blamed for recent attacks in the Iranian city of Rask in the southeastern province of Sistan-Baluchestan.

A reliable yet inexpensive tool for the estimation of flood water spread is conducive for efficient disaster management. The application of optical and SAR imagery in tandem provides a means of extended availability and enhanced reliability of flood mapping. We propose a methodology to merge these two types of imagery into a common data space and demonstrate its use in the identification of affected populations and infrastructure for the 2022 floods in Pakistan. The merging of optical and SAR data provides us with improved observations in cloud-prone regions; that is then used to gain additional insights into flood mapping applications. The use of open source datasets from WorldPop and OSM for population and roads respectively makes the exercise globally replicable. The integration of flood maps with spatial data on population and infrastructure facilitates informed policy design. We have shown that within the top five flood-affected districts in Sindh province, Pakistan, the affected population accounts for 31 %, while the length of affected roads measures 1410.25 km out of a total of 7537.96 km.