Inference optimizations such as quantization, pruning, format and datatype conversion, model export, and serialization can lead to functional degradations in language model task performance. While most efforts on performance recovery for deployment focus on robust quantization techniques, we focus on recovering model accuracies from any sources that degrade model weights, such as improper model serialization. In this work, we propose Recover-LoRA, a lightweight and dataset agnostic method to recover accuracy in degraded models. Recover-LoRA uses synthetic data and logit distillation to learn LoRA adapters on selective layers that facilitate aligning the degraded model to its full precision model. We investigate the utility of Recover-LoRA across a diverse set of small language models (SLMs), including models with varying attention architectures, multi-head attention (MHA) and group-query attention (GQA), as well as several evaluation datasets. Our results show that Recover-LoRA recovers model accuracies by 5-17% on MHA and GQA SLMs.

Reference-free metrics like self-perplexity are strongly biased against creative text generation. We propose the Confidence Score (CS), derived from a model's output probability distribution, as a less biased alternative. Experiments on gpt-4o-mini show that while fluency-based metrics prefer novel responses in 0\% of cases on 99 creative prompts, our CS does so 19% of the time, a statistically significant difference (95% CI for difference: [11.1%, 27.3%]). We also show that CS effectively distinguishes between easy, medium, and hard tasks, confirmed by non-overlapping confidence intervals. The Confidence Score thus mitigates the creativity bias of traditional metrics while retaining their core evaluative strengths, offering a more balanced assessment for modern LLMs.

A central question in artificial intelligence is the extent to which machine learning models comprehend mathematics. To address this, we propose a novel framework for measuring mathematical reasoning that moves beyond standard benchmarks to diagnose specific failure points. Our method first generates structured, step-by-step reasoning from gpt-3.5-turbo on the GSM8K dataset. We then use a more capable analyst model, gpt-4o-mini, to categorize errors and, crucially, perform an unsupervised clustering of every reasoning sentence to identify emergent "reasoning modes." This analysis reveals a cognitive profile with a stark, nonhuman-like brittleness: while the model achieves near-perfect accuracy on procedural modes like sequential calculation, its performance on modes requiring combinatorial reasoning with restrictions plummets. By identifying and quantifying the reliability of these distinct reasoning skills, our work provides a more granular method to evaluate mathematical comprehension and offers a precise roadmap for developing new capabilities and more reliable future applications.

Recent advancements in QML and SNNs have generated considerable excitement, promising exponential speedups and brain-like energy efficiency to revolutionize AI. However, this paper argues that they are unlikely to displace DNNs in the near term. QML struggles with adapting backpropagation due to unitary constraints, measurement-induced state collapse, barren plateaus, and high measurement overheads, exacerbated by the limitations of current noisy intermediate-scale quantum hardware, overfitting risks due to underdeveloped regularization techniques, and a fundamental misalignment with machine learning's generalization. SNNs face restricted representational bandwidth, struggling with long-range dependencies and semantic encoding in language tasks due to their discrete, spike-based processing. Furthermore, the goal of faithfully emulating the brain might impose inherent inefficiencies like cognitive biases, limited working memory, and slow learning speeds. Even their touted energy-efficient advantages are overstated; optimized DNNs with quantization can outperform SNNs in energy costs under realistic conditions. Finally, SNN training incurs high computational overhead from temporal unfolding. In contrast, DNNs leverage efficient backpropagation, robust regularization, and innovations in LRMs that shift scaling to inference-time compute, enabling self-improvement via RL and search algorithms like MCTS while mitigating data scarcity. This superiority is evidenced by recent models such as xAI's Grok-4 Heavy, which advances SOTA performance, and gpt-oss-120b, which surpasses or approaches the performance of leading industry models despite its modest 120-billion-parameter size deployable on a single 80GB GPU. Furthermore, specialized ASICs amplify these efficiency gains. Ultimately, QML and SNNs may serve niche hybrid roles, but DNNs remain the dominant, practical paradigm for AI advancement.

Generative AI (GenAI) has achieved remarkable success across a range of domains, but its capabilities remain constrained to statistical models of finite training sets and learning based on local gradient signals. This often results in artifacts that are more derivative than genuinely generative. In contrast, Evolutionary Computation (EC) offers a search-driven pathway to greater diversity and creativity, expanding generative capabilities by exploring uncharted solution spaces beyond the limits of available data. This work establishes a fundamental connection between EC and GenAI, redefining EC as Natural Generative AI (NatGenAI) -- a generative paradigm governed by exploratory search under natural selection. We demonstrate that classical EC with parent-centric operators mirrors conventional GenAI, while disruptive operators enable structured evolutionary leaps, often within just a few generations, to generate out-of-distribution artifacts. Moreover, the methods of evolutionary multitasking provide an unparalleled means of integrating disruptive EC (with cross-domain recombination of evolved features) and moderated selection mechanisms (allowing novel solutions to survive), thereby fostering sustained innovation. By reframing EC as NatGenAI, we emphasize structured disruption and selection pressure moderation as essential drivers of creativity. This perspective extends the generative paradigm beyond conventional boundaries and positions EC as crucial to advancing exploratory design, innovation, scientific discovery, and open-ended generation in the GenAI era.

The field of object detection and understanding is rapidly evolving, driven by advances in both traditional CNN-based models and emerging multi-modal large language models (LLMs). While CNNs like ResNet and YOLO remain highly effective for image-based tasks, recent transformer-based LLMs introduce new capabilities such as dynamic context reasoning, language-guided prompts, and holistic scene understanding. However, when used out-of-the-box, the full potential of LLMs remains underexploited, often resulting in suboptimal performance on specialized visual tasks. In this work, we conduct a comprehensive comparison of fine-tuned traditional CNNs, zero-shot pre-trained multi-modal LLMs, and fine-tuned multi-modal LLMs on the challenging task of artificial text overlay detection in images. A key contribution of our study is demonstrating that LLMs can be effectively fine-tuned on very limited data (fewer than 1,000 images) to achieve up to 36% accuracy improvement, matching or surpassing CNN-based baselines that typically require orders of magnitude more data. By exploring how language-guided models can be adapted for precise visual understanding with minimal supervision, our work contributes to the broader effort of bridging vision and language, offering novel insights into efficient cross-modal learning strategies. These findings highlight the adaptability and data efficiency of LLM-based approaches for real-world object detection tasks and provide actionable guidance for applying multi-modal transformers in low-resource visual environments. To support continued progress in this area, we have made the code used to fine-tune the models available in our GitHub, enabling future improvements and reuse in related applications.

Biomedical Named Entity Recognition (BioNER), task6 in BioASQ (A challenge in large-scale biomedical semantic indexing and question answering), is crucial for extracting information from scientific literature but faces hurdles such as distinguishing between similar entity types like genes and chemicals. This study evaluates the GLiNER-BioMed model on a BioASQ dataset and introduces a targeted dictionary-based post-processing strategy to address common misclassifications. While this post-processing approach demonstrated notable improvement on our development set, increasing the micro F1-score from a baseline of 0.79 to 0.83, this enhancement did not generalize to the blind test set, where the post-processed model achieved a micro F1-score of 0.77 compared to the baselines 0.79. We also discuss insights gained from exploring alternative methodologies, including Conditional Random Fields. This work highlights the potential of dictionary-based refinement for pre-trained BioNER models but underscores the critical challenge of overfitting to development data and the necessity of ensuring robust generalization for real-world applicability.

Abstract--Detecting psychological stress from speech is critical in high-pressure settings. While prior work has leveraged acoustic features for stress detection, most treat stress as a static label. In this work, we model stress as a temporally evolving phenomenon influenced by historical emotional state. We propose a dynamic labelling strategy that derives fine-grained stress annotations from emotional labels and introduce cross-attention-based sequential models--a Unidirectional LSTM and a Transformer Encoder--to capture temporal stress progression. Our approach achieves notable accuracy gains on MuSE (+5%) and StressID (+18%) over existing baselines, and generalises well to a custom real-world dataset. These results highlight the value of modelling stress as a dynamic construct in speech.

ABSTRACT Prior works have investigated the use of articulatory features as complementary representations for automatic speech recognition (ASR), but their use was largely confined to shallow acoustic models. In this work, we revisit articulatory information in the era of deep learning and propose a framework that leverages articulatory representations both as an auxiliary task and as a pseudo-input to the recognition model. Specifically, we employ speech inversion as an auxiliary prediction task, and the predicted articulatory features are injected into the model as a query stream in a cross-attention module with acoustic embeddings as keys and values. Experiments on LibriSpeech demonstrate that our approach yields consistent improvements over strong transformer-based baselines, particularly under low-resource conditions. These findings suggest that articulatory features, once sidelined in ASR research, can provide meaningful benefits when reintroduced with modern architectures.

Alzheimer's disease (AD) presents a complex, multifaceted challenge to patients, caregivers, and the healthcare system, necessitating integrated and dynamic support solutions. While artificial intelligence (AI) offers promising avenues for intervention, current applications are often siloed, addressing singular aspects of the disease such as diagnostics or caregiver support without systemic integration. This paper proposes a novel methodological framework for a comprehensive, multi-agent system (MAS) designed for holistic Alzheimer's disease management. The objective is to detail the architecture of a collaborative ecosystem of specialized AI agents, each engineered to address a distinct challenge in the AD care continuum, from caregiver support and multimodal data analysis to automated research and clinical data interpretation. The proposed framework is composed of eight specialized, interoperable agents. These agents are categorized by function: (1) Caregiver and Patient Support, (2) Data Analysis and Research, and (3) Advanced Multimodal Workflows. The methodology details the technical architecture of each agent, leveraging a suite of advanced technologies including large language models (LLMs) such as GPT-4o and Gemini, multi-agent orchestration frameworks, Retrieval-Augmented Generation (RAG) for evidence-grounded responses, and specialized tools for web scraping, multimodal data processing, and in-memory database querying. This paper presents a detailed architectural blueprint for an integrated AI ecosystem for AD care. By moving beyond single-purpose tools to a collaborative, multi-agent paradigm, this framework establishes a foundation for developing more adaptive, personalized, and proactive solutions. This methodological approach aims to pave the way for future systems capable of synthesizing diverse data streams to improve patient outcomes and reduce caregiver burden.