default author photo

Alessio Rocchi

AIgen Solutions

Head of AI Research

Via Dei Maffei 42

Rome, Rome 00165

Italy

http://aigensolutions.it

SCHOLARLY PAPERS

6

DOWNLOADS

219

TOTAL CITATIONS

0

Ideas:
“  My work is shifting toward the theory of composition in agentic systems: which formal objects let bounded reasoning units combine into systems you can analyze rather than only tune. Two directions sit at the center. The first is cognitive kernels, treating a reasoning primitive as a kernel in the analytic sense, so that composition and transport between reasoning states have a precise algebra instead of an ad hoc prompt chain. The second is mixture-of-experts orchestration, where routing and gating are the actual decision problem, and where the multi-agent and MoE views of specialization read as the same routing geometry seen twice. Both feed back into the applied work: LLM coding agents under hard constraints, retrieval versus long context in production RAG, and reinforcement learning for execution in algorithmic trading. The constant is a preference for structure over heuristics, and for reporting where the structure breaks.  ”

Scholarly Papers (6)

1.

Streamlined Hierarchical Reinforcement Learning for Algorithmic Trading: Architecture Simplification and Empirical Validation

Number of pages: 42 Posted: 07 Oct 2025 Last Revised: 08 Oct 2025
Alessio Rocchi
AIgen Solutions
Downloads 119 (621,896)

Abstract:

Loading...

Hierarchical Reinforcement Learning, Algorithmic Trading, Market Regime Detection, LSTM, Architecture Simplication, Empirical Validation, Mean Reversion, Financial Machine Learning

2.

Feature Scope and Cross-Sectional Return Prediction: Evidence from US Large-Cap Equities

Number of pages: 98 Posted: 27 Apr 2026
Alessio Rocchi
AIgen Solutions
Downloads 44 (1,448,301)

Abstract:

Loading...

3.

Conformance, Cost, and Replication in Constrained LLM Coding Agents

Number of pages: 32 Posted: 27 May 2026
Alessio Rocchi
AIgen Solutions
Downloads 20 (1,491,251)

Abstract:

Loading...

LLM Code Generation, Procedural Constraints, Stochastic Dominance, Asymmetric Review, Error Decorrelation, Predictability–Cost Trade-Off, AST-Level Mutation Testing, Reviewer Family Variation

4.
Downloads 15

Abstract:

Loading...

5.

FLEX-MoE: Failure-Born Orthogonal Experts for Self-Improving Language Models

Number of pages: 22 Posted: 16 Jul 2026
Alessio Rocchi
AIgen Solutions
Downloads 12

Abstract:

Loading...

Mixture of Experts, Low-Rank Adaptation (LoRA), Parameter-Efficient Fine-Tuning, Self-Improving Language Models, Continual Learning, Catastrophic Forgetting, Orthogonal Subspace Projection, Modular Adapters, Verifier-Grounded Learning, Safe Deployment, Large Language Models

Abstract:

Loading...