Natural Language Processing - Weekly Roundup

Edge service orchestration can achieve up to 64.5% lower decision latency than LLMs by using Jev decision models, even at high loads where LLMs fail.

This work studies a lightweight coding harness whose execution loop is fixed while three components are varied: planning, action space, and context management, and finds that context management becomes increasingly valuable as the context-window budget tightens.

Relational BabyLM is a system submission to the BabyLM 2026 challenge that combines two cognitively motivated inductive biases in a single decoder-only Transformer, and introduces a novel symbol-retrieval mechanism that matches learned symbol libraries while adding no parameters.

This work demonstrates how rigorous analytical frameworks can address complex humanitarian operations challenges characterised by information asymmetry and limited ground-truth data.

Astronex-World 1.0 is presented, an open controllable video world-model foundation that provides a bidirectional model for full-context generation and a causal model with block-causal attention and cross-block KV caching for persistent generation, both built on the Wan2.2-TI2V-5B prior.

Edge service orchestration can achieve up to 64.5% lower decision latency than LLMs by using Jev decision models, even at high loads where LLMs fail.

This work studies a lightweight coding harness whose execution loop is fixed while three components are varied: planning, action space, and context management, and finds that context management becomes increasingly valuable as the context-window budget tightens.

Relational BabyLM is a system submission to the BabyLM 2026 challenge that combines two cognitively motivated inductive biases in a single decoder-only Transformer, and introduces a novel symbol-retrieval mechanism that matches learned symbol libraries while adding no parameters.

This work demonstrates how rigorous analytical frameworks can address complex humanitarian operations challenges characterised by information asymmetry and limited ground-truth data.

Astronex-World 1.0 is presented, an open controllable video world-model foundation that provides a bidirectional model for full-context generation and a causal model with block-causal attention and cross-block KV caching for persistent generation, both built on the Wan2.2-TI2V-5B prior.

These cross-lingual transfer experiments show that linguistic proximity has no consistent predictive power for transfer performance, whereas structural similarity in rhetorical organization shows a weak but consistent positive correlation across models, and a set of three methods that explicitly leverage structural information using generative models are proposed.

CliniCIRCA is the first to temporally classify clinical events across unstructured discharge summaries without event-level timestamps and instruction tuning generally improves five open-weight models on event extraction, temporal tagging, and summarization across silver and clinician-verified evaluations.

A reusable three-layer checklist-and-judge framework that evaluates interpreter-mediated conversation across semantic, pragmatic, and cultural-social dimensions, covering the naturalness, intent, and social appropriateness that fidelity metrics leave unmeasured is introduced.

A closed-loop control architecture for a non-deterministic language model that must satisfy a numeric output constraint, that is, a requirement expressible as a number or an interval and checkable by code, is presented and evaluated.

The ANVIL compiler architecture is proposed, where ANVIL performs exceptionally well on leading benchmarks by using language models purely where they are effective, rather than as a catch-all tool.

The main finding is that evidence grounding is limited by retrieval, not by reading: given a research question, retrieve the relevant papers from a pool of 27,487, cite the page and the table or figure where the answer lives, and answer in a requested format.

It is hypothesised that contextual embeddings can complement frequency dynamics to effectively track diachronic semantic change, and frequency and embedding-based approaches across Astrophysics and NLP corpora spanning from 2010 to 2024 are compared.

CUEES (Correlation-gUided Encoder Selection), a lightweight heuristic that estimates complementarity through task- and category-level Pearson correlations between encoders'performance profiles, scoring a candidate set from single-encoder evaluations alone--without fusion training during selection.

An end-to-end computational pipeline for classifying intergovernmental agreements by institutional form and extracting financial relationships between principals and agents in service contracts is introduced, rendering the content of Iowa interlocal agreements analyzable at scale for the first time.

A Gated Noise-filtered Sentiment-Relevance Interaction (GNSRI) framework that employs a gated noise-filtering module to suppress sentiment-irrelevant features and enhance aspect-aware sentiment cues, and a sentiment-relevance interaction module to capture consistent and conflicting cross-modal signals at micro and macro levels.

Token Importance-Aware Policy Optimization is proposed, a novel reinforcement learning strategy that explicitly leverages token-importance awareness and identifies core tokens based on token dependency and reweights a trajectory's advantage according to its overall dependencies.

Results show mechanism-level evaluation captures architecture-specific sensitivity patterns that behavioral benchmarks cannot; pairing both should become standard NLP practice.

Findings indicate that reliable context reduction depends more on preserving protocol-critical state than on maximizing token removal, and that adaptive guardrails can improve efficiency, scalability, and reliability in long-horizon agentic systems.

This work explores how diffusion can enable multilingual translation with a discrete diffusion framework that refines all target languages in parallel, achieving sublinear latency scaling with the number of targets, and supports deployment as a single unified model to replace multiple independent systems.

A matched ablation on unfiltered training data shows substantial benefits from filtering for all fine-tuned models, with the LLM-based models gaining roughly twice as much from filtering as the dedicated encoder-decoder MT system, supporting the view that for Belarusian MT one of the primary bottlenecks is data quality.

StalePO is introduced, an objective derived from three requirements this regime imposes that improves the fraction of segments passing all LLM-as-judge MQM quality checks by 14.9 and 4.6 percentage points, respectively, with gains concentrated on style and fluency.

This paper investigates how AI-driven feedback mechanisms, including sentiment analysis systems, pulse surveys, chatbots, engagement dashboards, and predictive analytics, are changing African workplaces through offering continuous listening, instant performance information and proactive engagement with employees.

A benchmark dataset specifically tailored to the ESG domain is constructed, comparing their performance with respect to retrieval, and Retrieval-Augmented Generation (RAG), and provides practical insights into which models are better suited for ESG RAG tasks.