№ 02 / SUMMARIES

#machine-learning

Every summary, chronological. Filter by category, tag, or source from the rail.

Tag · #machine-learning
DAY 01Today AUG 29 · 20269 SUMMARIES
arXiv cs.AIAI & LLMs

Explainable AI Frameworks for Telecom Churn Prediction

This paper proposes a framework for integrating Explainable AI (XAI) into CRM systems to improve the transparency and actionability of customer churn predictions in telecommunications.

arXiv cs.AI
arXiv cs.AIAI & LLMs

The Accuracy-Efficiency Paradox in On-Device Energy Forecasting

On-device energy forecasting models often consume more power than the energy savings they aim to provide, creating a net-negative efficiency paradox that requires careful calibration of model complexity.

arXiv cs.AIAI & LLMs

EEG-to-Report: Bridging Clinical Brain Data and Language Models

The EEG-to-Report framework introduces a standardized annotation and feature-text mapping method to enable training language models on complex clinical EEG data, bridging the gap between raw neural signals and diagnostic reports.

arXiv cs.AIAI & LLMs

Building Safe Multimodal AI for Mental Health Support

The Anian framework introduces a safety-gated architecture for mental health AI, utilizing hierarchical state representation and conservative risk fusion to ensure controlled, reliable patient interactions.

arXiv cs.AIAI & LLMs

Refusal Is Not Robustness: LLMs Fabricate on Uninformative Data

Large Language Models often fail to identify uninformative input, choosing to confidently fabricate clinical assessments rather than admitting a lack of sufficient data.

arXiv cs.AIData Science & Visualization

The 5D Framework for Multi-Table Data Analysis

The 5D framework provides a unified methodology for integrating and reusing complex, multi-table datasets by mapping data across five distinct dimensions to ensure consistency and analytical depth.

arXiv cs.AIAI & LLMs

Explaining ICU Mortality Predictions with LLM Agentic Pipelines

This study demonstrates the feasibility of using standalone LLMs and pre-specified agentic pipelines to interpret complex ICU mortality risk models, providing a path toward more transparent clinical decision support.

arXiv cs.AIAI & LLMs

EduRiskX: Combining Transformers and F-Logic for Academic Prediction

EduRiskX improves academic risk prediction by pairing temporal Transformers for pattern recognition with F-Logic for rule-based, interpretable reasoning.

arXiv cs.AIAI & LLMs

CIFQA: Deterministic Multi-Agent Framework for Financial Analysis

CIFQA is a multi-agent framework designed to improve financial query accuracy by replacing non-deterministic LLM reasoning with a structured, tool-grounded execution pipeline.

DAY 02Thursday AUG 27 · 20268 SUMMARIES
IBM TechnologyAI & LLMs

Beyond Leaderboards: Evaluating Real-World AI Systems

Model benchmarks are just a starting point; production reliability requires balancing accuracy, latency, and cost through system-level evaluations and agentic chain testing.

IBM Technology
arXiv cs.AIAI & LLMs

A Formal Framework for Auditing XAI Robustness and Fidelity

This paper proposes a formal methodology to audit Explainable AI (XAI) systems, ensuring that explanations are both robust to input perturbations and faithful to the underlying model's decision-making process.

arXiv cs.AIAI & LLMs

Decoupling Model Performance from Evaluation Bias

Current AI benchmarks often conflate model capability with the biases of the evaluation instrument itself, necessitating a shift toward disentangling model preferences from measurement artifacts.

arXiv cs.AIAI & LLMs

Optimizing Code Models with Function-Level Execution Feedback

Improving code generation models by using granular, function-level execution feedback rather than binary pass/fail signals to guide preference optimization.

arXiv cs.AIAI & LLMs

RL-Enhanced Agentic Search for Biomedical Fact-Checking

This paper introduces a reinforcement learning-based agentic framework designed to improve the accuracy and reliability of automated biomedical fact-checking by optimizing search strategies.

arXiv cs.AIAI & LLMs

Reducing Medical AI Sycophancy via Gated Activation Steering

Gated Activation Steering (GAS) improves medical LLM reliability by dynamically suppressing internal representations associated with sycophancy and hallucinations during inference, without requiring model retraining.

arXiv cs.AIAI & LLMs

Optimizing Masked Diffusion LLMs for Real-World Hardware

This paper provides a characterization of Masked Diffusion LLMs, identifying unique computational bottlenecks and proposing hardware-aware design principles to improve inference efficiency.

arXiv cs.AIAI & LLMs

RENDER: A Framework for Controlling Evidence in LLM Memory Evaluation

RENDER is a new evaluation framework designed to isolate and measure how LLMs process and recall specific evidence within their context windows, addressing the limitations of existing memory benchmarks.

DAY 03Wednesday AUG 26 · 20268 SUMMARIES
Google Cloud TechSoftware Engineering

Strategies for Serving JAX Models in Production

Moving JAX models from notebooks to production requires choosing the right serialization and compilation strategy to avoid latency spikes caused by just-in-time compilation.

Google Cloud Tech
Google Cloud TechSoftware Engineering

Scaling JAX Models to Multi-GPU Systems

Scale JAX models across multiple GPUs by defining array layouts with Mesh and PartitionSpec, allowing the compiler to handle gradient synchronization automatically.

Google Cloud TechSoftware Engineering

Building and Optimizing JAX Training Loops

Build high-performance JAX training loops by maintaining pure functions, keeping data on-device, and utilizing fused kernels like cuDNN attention to avoid GPU memory bottlenecks.

Google Cloud TechAI & LLMs

Optimizing JAX Performance on NVIDIA GPUs

JAX performance hinges on ensuring your code runs on the GPU, maintaining stable input shapes to prevent re-compilation, and correctly handling asynchronous execution during profiling.

arXiv cs.AIAI & LLMs

Architecture-Aware Credit Transport for LLM Reinforcement Learning

The paper introduces a method to improve LLM reinforcement learning by aligning credit assignment with the underlying computational architecture, ensuring rewards are distributed based on actual processing paths.

arXiv cs.AIAI & LLMs

Hate Speech Classification in Roman Urdu: PEFT vs. Prompt Engineering

A comparative study evaluating Parameter-Efficient Fine-Tuning (PEFT) against prompt engineering for detecting hate speech in Roman Urdu, highlighting the trade-offs between computational efficiency and classification accuracy in low-resource linguistic contexts.

arXiv cs.AIData Science & Visualization

Frameworks for Explainable AI in Time Series Classification

A systematic review of current software frameworks for XAI in time series classification, highlighting the need for standardized evaluation and better integration of interpretability tools in production pipelines.

arXiv cs.AIAI & LLMs

KVBoost: Accelerating LLM Inference via Chunk-Level Cache Reuse

KVBoost improves LLM inference latency by 4.49x by enabling chunk-level KV cache reuse regardless of position, using a dual-hash keying scheme and deviation-guided recomputation to maintain accuracy.

DAY 04Tuesday AUG 25 · 20265 SUMMARIES
arXiv cs.AIAI & LLMs

Structurally Indirect Prerequisite Eviction in Agentic Memory

Agentic memory systems often fail not due to retrieval errors, but because 'prerequisite' information is evicted from context before it can be used, creating a structural failure in long-term reasoning.

arXiv cs.AI
arXiv cs.AIAI & LLMs

Defining World Models for Agents and Environments

This paper provides a formal framework for understanding world models by distinguishing between environment-only, agent-only, and joint agent-environment system dynamics.

arXiv cs.AIAI & LLMs

Foundations and Frontiers of Multimodal Agentic Frameworks

Multimodal agentic frameworks integrate diverse sensory inputs with reasoning capabilities, moving beyond text-only models to enable autonomous task execution in complex, real-world environments.

arXiv cs.AIAI & LLMs

Representation Affects Retrieval in Multimodal Agent Routing

The effectiveness of skill discovery and routing in multimodal agents is fundamentally constrained by the quality of the underlying data representation, proving that retrieval performance is inseparable from how skills are encoded.

arXiv cs.AIData Science & Visualization

Interpretable Multimodal Classification via Linear Discriminant Trees

The paper proposes Linear Discriminant Tree Ensembles (LDTE) as a method to achieve high-accuracy multimodal classification while maintaining model interpretability through hierarchical linear decision boundaries.

Showing 30 of 539