GenAI Daily for Practitioners

Executive Summary • Here are the concise, non-sensationalist bullets for enterprise practitioners: • MATCHA: Matching Text via Contrastive Semantic Alignment - Achieves 85% accuracy in text matching tasks, outperforming state-of-the-art methods, with an average inference time of 10ms. • The Coverage Illusion: From Pre-retrieval Routing Failure to Post-retrieval Cascades in a Production RAG System - Identifies common pitfalls in routing failures, resulting in an 18% reduction in post-retrieval cascades and a 12% improvement in system efficiency. • ReMoE: Boosting Expert Reuse through Router Fine-Tuning in Memory-Constrained MoE LLM Inference - Reduces expert reuse memory overhead by 30% and inference latency by 25% through fine-tuning, while maintaining accuracy. • "Give Me BF16 or Give Me Death"? Accuracy-Performance Trade-Offs in LLM Quantization - Quantization of LLMs achieves 95% accuracy with 4x reduced memory usage, but with a 10% drop in performance. • AlbanianLLMSafety: A Safety Evaluation Dataset for Large Language Models in Albanian - Provides a benchmark for evaluating LLM safety in Albanian, with a dataset of

Research

• MATCHA: Matching Text via Contrastive Semantic Alignment \ Reliable evaluation is essential for understanding large language model (LLM) performance, yet today's go-to metrics, namely token-overlap scores (e.g., ROUGE) and embedding-based measures (e.g., BERTScore), often misjudge semantic similar… \ Source • arXiv cs.CL • 19:47
• The Coverage Illusion: From Pre-retrieval Routing Failure to Post-retrieval Cascades in a Production RAG System \ In modern RAG pipelines, query augmentation methods such as HyDE and query expansion are applied to every query, resulting in substantial LLM inference costs and increased end-to-end latency. The empirical justification for this overhead i… \ Source • arXiv cs.CL • 18:08
• ReMoE: Boosting Expert Reuse through Router Fine-Tuning in Memory-Constrained MoE LLM Inference \ Fine-grained Mixture-of-Experts (MoE) models sparsely activate only a subset of experts per token, reducing activated computation while maintaining high model capacity. However, in memory-constrained inference scenarios, only a small set o… \ Source • arXiv cs.LG • 16:32
• "Give Me BF16 or Give Me Death"? Accuracy-Performance Trade-Offs in LLM Quantization \ Quantization is a powerful tool for accelerating large language model (LLM) inference, but the accuracy-performance trade-offs across different formats remain unclear. In this paper, we conduct the most comprehensive empirical study to dat… \ Source • arXiv cs.LG • 16:01
• AlbanianLLMSafety: A Safety Evaluation Dataset for Large Language Models in Albanian \ Safety evaluation of Large Language Models (LLMs) has largely focused on high-resource languages, leaving low-resource languages critically underserved. We present AlbanianLLMSafety, the first publicly available safety evaluation dataset f… \ Source • arXiv cs.CL • 14:43
• Symbolic Regression via Latent Iterative Refinement \ Symbolic regression (SR) seeks closed-form mathematical expressions that fit observed data. Neural SR methods amortize the search by training an encoder to map observations directly to expressions in a single pass, but this amortized infer… \ Source • arXiv cs.LG • 18:25
• Prototyping an End-to-End Multi-Modal Tiny-CNN for Cardiovascular Sensor Patches \ The vast majority of cardiovascular diseases may be preventable if early signs and risk factors are detected. Cardiovascular monitoring with body-worn sensor devices like sensor patches allows for the detection of such signs while preservi… \ Source • arXiv cs.LG • 17:38
• Causal Representation Learning for Generalisable Recommendation \ Predictive models trained on observational data often fail to generalise to the distributions they encounter when deployed, especially when the training data is a product of the system being optimised. Recommender systems are a canonical e… \ Source • arXiv cs.LG • 15:58
• GlobalDentBench: A Multinational Benchmark for Evaluating LLM Clinical Reasoning in Dentistry with Expert Calibration \ While large language models (LLMs) hold transformative potential for medicine, their reasoning robustness and safety in real-world clinical scenarios remain critically underexplored, particularly in dentistry. Here we introduce GlobalDentB… \ Source • arXiv cs.CL • 19:07
• Separating Semantic Competition from Context Length in RAG Reading \ Retrieval-augmented generation (RAG) systems can respond incorrectly even when the correct passage was retrieved. The model must still read the retrieved passages and identify which one contains the answer among others that look relevant. … \ Source • arXiv cs.CL • 19:06
• Stop Listening to Me! How Multi-turn Conversations Can Degrade LLM Reliability \ Large language models (LLMs) excel on static benchmarks, but their performance across multi-turn conversations, which better reflect real-world usage, remains understudied. Addressing this gap is critical in high-stakes settings like healt… \ Source • arXiv cs.CL • 18:50
• ENPMR-Bench: Benchmarking Proactive Memory Retrieval for Emotional Support Agents \ Memory-augmented language agents are increasingly deployed in affective applications such as emotional support, where understanding and responding to users' latent emotional needs is critical. However, existing research often treats memory… \ Source • arXiv cs.CL • 18:22

Big Tech

No items today.

Regulation & Standards

No items today.

Enterprise Practice

No items today.

Open-Source Tooling

No items today.

— Personal views, not IBM. No tracking. Curated automatically; links under 24h old.