#evaluation optimization separation
72 related articles
·4 min
Sliced Wasserstein Distance for Improving Class Separability: Experiments and Reflections
Exploring experiments using Sliced Wasserstein Distance (SWD) to learn feature transformations that increase inter-class distribution distance. Analyzing why this approach works for decision trees but fails for other classifiers.
Read more →
·3 min
Generate Xianxia Wallpapers with a Single Sentence: A Hands-On Comparison of Three AI Agents with Skill Enhancement
Testing the "Eastern Xianxia Visual Director" Skill across Codex, WorkBody, and Grog to see how a single plain sentence becomes stunning xianxia wallpaper art.
Read more →
·4 min
Reframing Low-Resource Machine Translation with POMDP: Designing a Bengali Translation Agent
Explore how POMDP remodels low-resource machine translation for Bengali, combining MBR decoding and active disambiguation to tackle ambiguity, code-mixing, and speech noise.
Read more →
·3 min
Where's the Value of ML Engineers in the Agent Era? The Shift from Executor to Problem Definer
As AI Agents independently handle training optimization, ML engineers must shift from code executors to problem definers—building tamper-proof evaluation systems and governing Agent behavior.
Read more →
·4 min
Dify in Practice: Enterprise-Grade End-to-End Pipeline Design for Natural Language to SQL
Build a complete NL2SQL solution on Dify with three knowledge bases, multi-model judge mechanism, SQL security validation, and ECharts visualization.
Read more →
·4 min
PySpark Crash Course: Core Principles of Distributed Data Processing & Five Key Use Cases
A systematic guide to PySpark's core architecture, lazy evaluation, and Shuffle mechanism, covering five practical scenarios: ETL pipelines, DataFrame vs RDD, streaming, and distributed ML.
Read more →
·6 min
Cloudflare OS Deep Dive: Security Architecture and Implementation Path of an Open-Source Enterprise Agent Platform
Deep dive into Cloudflare OS open-source enterprise agent platform, covering zero-permission security model, Gatekeeper governance, agent workspaces, application architecture, and model-agnostic strategy.
Read more →
·3 min
34 Model Iterations in Review: Most Problems Were Actually in the Evaluation Pipeline
After 34 model iterations, an AIOps engineer found most gains came from evaluation bugs. This article details three critical evaluation pitfalls and solutions for MLOps practitioners.
Read more →
·4 min
6 Must-Do Tasks Before Deploying Enterprise AI Agents: A Complete Guide to Observability + Evaluation
A detailed guide to 6 critical engineering challenges for enterprise AI Agents before production, covering Langfuse-based tracing, observability, evaluation stages, prompt governance, and high-concurrency architecture.
Read more →
·4 min
Total Parameter Count Is Obsolete: Understanding the Two-Number Era of MoE Models
Mixture-of-Experts (MoE) makes single parameter counts obsolete. Learn the difference between total and active parameters, how MoE decouples knowledge capacity from inference cost, and why this matters for model selection.
Read more →
·3 min
ACAI Architecture Explained: How Modular Cognitive Pipelines Solve LLM Hallucination and Context Corruption
Deep dive into the ACAI (Adaptive Cognitive AI) modular architecture that solves LLM hallucination and context window rot through layered cognitive pipelines, semantic memory graphs, and logical verification.
Read more →
·3 min
For Anxious ML Newcomers: How to Escape the Comparison Trap
Facing ML's rapid iteration and social media's survivorship bias, many newcomers fall into self-doubt. This article offers practical advice for escaping the comparison trap and rebuilding self-efficacy.
Read more →
·4 min
From Zero to AI Engineer: A Real Growth Path from a Student's Resume
A CS student went from Python basics to model deployment in 3-4 months, building an AI portfolio through three real projects. This article breaks down the learning path, project value, and resume optimization strategies.
Read more →
·4 min
Walsh Multi-Agent Trading System: How Risk Management Veto Power Ensures AI Autonomous Decision-Making Safety
Deep dive into the Walsh multi-agent trading system architecture, exploring how its risk management agent with veto power establishes safety boundaries for AI autonomous decision-making.
Read more →
·3 min
Implementing Zanzibar in Lean4: A Formally Verified Permission Engine for AI Projects
Deep dive into a Datalog permission DSL built on Google Zanzibar using Lean4 theorem prover. How formal verification strengthens AI permission management.
Read more →
·3 min
Why Are Uncensored Open-Source LLMs More Optimistic? A Deep Dive into Alignment Side Effects
Research finds uncensored open-source LLMs are measurably more optimistic than base models. This article analyzes how uncensoring changes model personality and the coupling effects of alignment.
Read more →
·5 min
July 24 AI Daily Brief: Flux 3 Multimodal Release, Kimi K3 Lags in Testing, Qwen Tops TTS Rankings
July 24 AI news: Black Forest Labs launches Flux 3 multimodal model, Kimi K3 lags in US-UK gov tests, Alibaba Qwen tops TTS rankings, Etched raises $300M, AMD unveils MI430X.
Read more →
·4 min
Deep Dive into the EYG Programming Language: A New Portable Programming Paradigm Designed for Humans
Deep analysis of the EYG programming language's core design, including algebraic effects, program state persistence, and cross-platform portability, exploring how it addresses modern software fragmentation.
Read more →
·3 min
Deconstructing Agents: Task Decomposition, Context Compression, and the Case Against Heavy Frameworks
A deep engineering analysis of Agent internals: how LLMs decompose tasks via tool calling, why context compression and memory are essential, and why solo developers should avoid heavy frameworks.
Read more →
·5 min
The Five Stages of AI Agent Architecture Evolution: The Real Path from Model Calls to Production Runtime
Deep dive into the five evolution stages of AI Agent architecture: model calls, tool calls, workflows, Agent loops, and production runtime. Learn the responsibility boundaries and design principles.
Read more →