Input Output Research (IOR), the research division of IOG (Input Output Group), published 24 peer-reviewed papers in 2025 as part of the first year of Cardano’s five-year plan, the “Cardano Vision.” Input Output Research (IOR)
In this post, we’ve analyzed all 24 papers with AI assistance and compiled clear summaries for each — organized by theme.
These papers could have a major impact on the future direction of Cardano, so let’s dive in!
- Consensus & Ouroboros
- Fairness & Fee Design
- Game Theory & Mechanism Design
- #08 Reward Schemes and Committee Sizes in Proof of Stake Governance
- #09 Pool Formation in Oceanic Games: Shapley Value and Proportional Sharing
- #10 Single-Token vs Two-Token Blockchain Tokenomics
- #11 Incentivizing Geographic Diversity for Decentralized Systems
- #12 On Sybil-proofness in Restaking Networks
- #13 Participatory Budgeting with Donations: The Case of Selective Voters
- #14 Accelerated Preference Elicitation with LLM-Based Proxies
- Zero-Knowledge Proofs & Cryptography
- Networking & Distributed Systems
- Rollups & Layer 2
- Hydra & Interchains
- Formal Methods & Programming Languages
- Summary by Categories
Consensus & Ouroboros
#01 Ouroboros Leios
📍 Crypto 2025 ／ 👤 Panagiotakos et al. ／ #consensus #scalability
A new consensus protocol designed to fundamentally solve the throughput limitations of Cardano’s Ouroboros Praos. The paper identifies a structural inefficiency: block diffusion in Praos occupies only about 25% of block time, leaving the remaining 75% idle. To address this, Leios introduces a three-tier parallel processing architecture consisting of “ranking blocks,” “input blocks,” and “endorsement blocks.” The goal is to increase throughput from the current ~4.4 kB/s to a theoretical maximum of ~900 kB/s. The protocol maintains security under up to 50% adversarial stake and includes a fallback mechanism to Praos. After its presentation at Crypto 2025, the work was formally handed over to Input Output Engineering (IOE) for implementation.
Related post: Ouroboros Leios — Everything About the “Input Endorser” That Will Maximize Cardano’s Throughput
#02 Adaptively Secure Fast Settlement with Dynamic Participation and Self-Healing
📍 December 2025 ／ 👤 Badertscher, Coretti, Gaži, Kiayias, Russell ／ #settlement #consensus
This paper provides the theoretical foundation for Ouroboros Peras — a protocol designed to significantly reduce settlement time after block production. The paper delivers formal proofs of safety, liveness, and self-healing. “Adaptively secure” means security is maintained even if an adversary dynamically corrupts participants during protocol execution, while “dynamic participation” means the protocol continues to operate normally even when SPOs go temporarily offline. The paper also demonstrates, through consultation with the engineering team, that integration with existing Praos requires minimal changes.
#03 Taming Iterative Grinding Attacks on Blockchain Beacons
📍 Asiacrypt 2025 ／ 👤 Gaži, Quader, Russell ／ #grinding #consensus
This paper addresses “grinding attacks” in PoS blockchains — where an adversary attempts to manipulate the randomness (nonce) used for epoch leader selection to their advantage. Cardano’s Praos has long faced this challenge. The paper rigorously analyzes the theoretical upper bound of iterative grinding attacks, providing the quantitative foundation for Ouroboros Φalanx (the anti-grinding CIP). Acceptance at Asiacrypt 2025, one of the top cryptography conferences, reflects the high quality of this research.
#04 Efficient and Proof-of-Useful-Work Friendly Local-Search for Distributed Consensus
📍 ePrint Archive, November 2025 ／ 👤 Fitzi, Kiayias, Michel, Panagiotakos, Russell ／ #proof-of-useful-work #consensus
This paper advances “Ofelimos,” IOG’s protocol introduced at Crypto 2022 that replaces the meaningless computational puzzles of traditional Proof-of-Work with useful combinatorial optimization problems. Ofelimos’s usefulness-to-wastefulness ratio depends heavily on the performance of its internal “Doubly-Parallel Local Search (DPLS)” algorithm. This paper designs a new, more efficient local search algorithm suited for distributed consensus environments. It is part of a broader research stream that could eventually lead to “Proof-of-Deep-Learning,” where deep learning inference computations serve as an alternative to PoW.
Fairness & Fee Design
#05 Universally Composable Transaction Order Fairness: Refined Definitions and Adaptive Security
📍 Asiacrypt 2025 ／ 👤 Ciampi, Kiayias, Shen ／ #order-fairness #universal-composition
This paper formally defines the question of whether transactions are processed in a fair order on a blockchain, using the Universal Composability (UC) framework. The manipulation of transaction ordering by block producers to extract profit — as exemplified by the MEV (Maximal Extractable Value) problem — is a serious issue in DeFi ecosystems. The paper refines the concept of “bounded unfairness” and establishes a formal definition for what degree of unfairness can be considered acceptable. It further proves that these definitions hold even under “adaptive security,” where an adversary can dynamically corrupt participants during execution.
#06 One-dimensional vs. Multi-dimensional Pricing in Blockchain Protocols
📍 WINE 2025 ／ 👤 Kiayias, Koutsoupias, Panagiotakos, Zioga ／ #game-theory #incentives #fees
This paper compares one-dimensional fee models (like Ethereum’s EIP-1559, which prices primarily gas) with multi-dimensional models that price multiple resources (compute, storage, bandwidth, etc.) separately. Cardano already uses a multi-dimensional fee structure that accounts for CPU and memory, and this research provides theoretical backing for that design. The paper formally proves that one-dimensional models cause non-inclusive price spikes during congestion, and analyzes the trade-offs of multi-dimensional models in detail. Reading this alongside the Tiered Mechanisms paper reveals the future direction of Cardano’s fee design.
#07 Tiered Mechanisms for Blockchain Transaction Fees
📍 WINE 2025 ／ 👤 Kiayias, Koutsoupias, Panagiotakos, Zioga ／ #game-theory #incentives
This paper designs and analyzes tiered fee mechanisms to allow blockchain DApps to serve diverse user segments — from urgent DeFi traders to cost-conscious everyday users. The paper formally proves that Ethereum’s EIP-1559 causes horizontal price surges during congestion that exclude low-budget users, a design flaw it terms “non-inclusive.” The tiered model sets different price tiers based on service priority, enabling a fairer fee environment where diverse users can coexist. This is directly relevant to Cardano’s fee design after the introduction of Leios.
Game Theory & Mechanism Design
#08 Reward Schemes and Committee Sizes in Proof of Stake Governance
📍 Financial Cryptography 2025 ／ 👤 Birmpas, Lazos, Markakis, Penna ／ #game-theory #governance #drep
Directly motivated by the DRep system introduced in Cardano’s CIP-1694, this paper provides a game-theoretic analysis of reward design and committee sizes in PoS governance. The model assumes that the more effort a DRep invests (information gathering costs), the higher their probability of voting correctly, which in turn attracts more delegation. The paper shows that proportional reward schemes lead to effort minimization at equilibrium, producing a low probability of selecting the correct outcome. A threshold mechanism — rewarding only DReps that exceed a certain delegation threshold — is shown to be superior. The paper also concludes that, from an optimization standpoint, small committees are sufficient. The assumptions, limitations, and real-world applicability of this paper warrant separate discussion.
Related post: What Is a DRep? A Complete Guide to Cardano Governance!
#09 Pool Formation in Oceanic Games: Shapley Value and Proportional Sharing
📍 AFT 2025 ／ 👤 Kiayias, Koutsoupias, Markakis, Tsamopoulos ／ #game-theory #stake-pools
This paper analyzes equilibria for pool participation in environments where an infinite number of small stakeholders (an “oceanic game”) coexist with a finite number of large players — with Cardano’s stake pool formation specifically in mind. It compares equilibria under two reward schemes: the Shapley value (a measure of fair distribution in cooperative game theory) and proportional sharing. Proportional sharing is closer to Cardano’s current approach, while Shapley value-based distribution more accurately rewards each participant’s contribution. The paper provides the theoretical foundation for the Shapley value-based reward distribution proposed in the TO-2 stream, with significant implications for decentralization.
#10 Single-Token vs Two-Token Blockchain Tokenomics
📍 AFT 2025 ／ 👤 Kiayias, Lazos, Penna ／ #tokenomics
This paper compares long-run equilibria between single-token designs (where users and validators use the same token, like ADA) and dual-token designs (where payment and staking tokens are separated) in PoS blockchains. Single-token designs risk falling into feedback loops where rewards grow exponentially in response to service value fluctuations. Dual-token designs, by contrast, can provide stable user token pricing while still offering adequate validator compensation. This paper provides the theoretical grounding for Cardano’s ADA-only design and forms a foundational framework for discussing future tokenomics reform.
#11 Incentivizing Geographic Diversity for Decentralized Systems
📍 Financial Cryptography 2026 ／ 👤 Roeschlin, Markakis, Bhaskar, Kiayias ／ #game-theory #incentives #decentralization
This paper proposes mechanism designs using economic incentives to counteract the tendency of participants in decentralized networks (such as SPOs and DReps) to concentrate geographically. Geographic concentration — whether in San Francisco, Tokyo, or elsewhere — weakens a network’s resilience to outages, censorship, and regulatory risk. By tying rewards to location through game-theoretic design, the paper seeks to create structures where participants benefit from spreading out. This research is directly relevant to SPO concentration in Cardano and to strengthening governance participation from the Asian and Japanese communities.
#12 On Sybil-proofness in Restaking Networks
📍 DeFi 2026 ／ 👤 Chitra, Penna, Schneider ／ #game-theory #sybil-proofness
This paper analyzes Sybil attack resistance in restaking networks — systems like EigenLayer where a single stake is reused as security collateral across multiple protocols. Sybil attacks, where a single actor creates multiple identities (wallets, nodes) to inflate rewards or voting power, are structurally identical to the DRep multi-wallet problem in Cardano. The paper shows that Sybil attacks in restaking environments have more complex game-theoretic structures than in standard settings, and formalizes the conditions needed to achieve Sybil-proofness. The findings are also relevant to security design in Cardano’s partner chains.
#13 Participatory Budgeting with Donations: The Case of Selective Voters
📍 ECAI 2025 ／ 👤 Lazos, Markakis, Papasotiropoulos ／ #social-choice #governance
This paper analyzes how to achieve fair and efficient budget allocation in participatory budgeting mechanisms when “selective voters” — who only participate in votes on a subset of projects — are present. It also accounts for the impact of external donations. The paper is positioned as a theoretical foundation for decentralized fund allocation systems like Cardano’s Project Catalyst, and may offer prescriptions for real-world challenges such as fund concentration in large projects and uneven participation rates.
#14 Accelerated Preference Elicitation with LLM-Based Proxies
📍 WINE 2025 ／ 👤 Huang, Lock, Marmolejo-Cossio, Parkes ／ #auctions #AI #governance
This paper proposes a method to accelerate the process of efficiently eliciting individual “preferences” from a large set of options using LLM (large language model)-based proxies. Normally, accurately gathering preference information requires many rounds of questions and answers. But LLM proxies can estimate and supplement preferences from past response patterns and context, enabling high-accuracy preference data to be obtained with fewer interactions. While the paper targets applications in governance and resource allocation decision support, it also highlights the light and shadow of Cardano governance participants using AI to make voting decisions.
Zero-Knowledge Proofs & Cryptography
#15 The Brave New World of Global Generic Groups and UC-Secure Zero-Overhead SNARKs
📍 TCC 2024 (listed on IOG in 2025) ／ 👤 Kohlweiss et al. ／ #zk-proofs #cryptography
This paper proves that SNARKs (Succinct Non-interactive ARguments of Knowledge) — as exemplified by Groth16 — can be realized under the Universal Composability (UC) security framework without any efficiency loss. As the term “zero-overhead” implies, the paper shows it is theoretically possible to achieve the strong property of UC security while maintaining the same proof size and verification cost as conventional SNARKs. This foundational research serves as the theoretical basis for Cardano’s Mithril and interchain research leveraging ZK proofs, directly contributing to future privacy-preserving features and scaling technologies.
#16 Efficient Batch Opening Schemes for Merkle Tree Commitment with Applications to Trustless Crosschain Bridge
📍 ICCCN 2025 ／ 👤 IC-1 stream authors ／ #zk-proofs #crosschain
This paper designs an efficient batch opening scheme for Merkle tree commitments and demonstrates its application to trustless cross-chain bridges. When a cross-chain bridge must “prove the state of one chain on another chain,” the computational cost is normally enormous. By using ZK-based batch Merkle openings, this cost can be drastically reduced. The paper is positioned as a technical pillar of interchain research (the IC-1 stream), including the Cardinal protocol that connects Bitcoin and Cardano.
Networking & Distributed Systems
#17 SCRamble: Adaptive Decentralized Overlay Construction for Blockchain Networks
📍 ICDCN 2026 ／ 👤 Kolyvas, Antonov, Voulgaris ／ #networking
This paper proposes the SCRamble protocol, which dynamically and decentrally optimizes the overlay (node connection topology) of blockchain networks. In P2P networks like Bitcoin and Cardano, connection structures tend to become rigid in response to node joins, departures, and changing network conditions, causing propagation delays and single points of failure. SCRamble enables nodes to autonomously reconstruct connections using only local information, maintaining global efficiency and fault tolerance. This is an important enabling technology for the large volumes of blocks and endorsement blocks expected to propagate through the network after Leios is introduced.
#18 Setchain Algorithms for Blockchain Scalability
📍 ACM CCS 2025 ／ 👤 Karmegam, Bianchi, Capretto, Ceresa, Fernandez Anta, Sanchez ／ #bft #scalability
This paper proposes a new distributed data structure called “Setchain” and its algorithms, applying them to improve blockchain scalability. In a “set”-based ledger model where total ordering of transactions is not required, the paper designs and evaluates multiple algorithms for efficient agreement in Byzantine Fault Tolerant (BFT) environments. Acceptance at ACM CCS (Computer and Communications Security), one of the top conferences in both security and systems, reflects the paper’s quality across both dimensions. The paper also provides theoretical foundations for potential integration with deterministic ledger systems like Cardano.
Rollups & Layer 2
#19 A Secure Sequencer and Data Availability Committee for Rollups (Extended Version)
📍 ACM CCS 2025 ／ 👤 Capretto, Ceresa, Fernandez Anta, Moreno-Sanchez ／ #rollups #formalmethods
This paper provides formal security definitions for two critical components of Layer 2 blockchain rollups — the “sequencer” (which determines transaction ordering) and the “data availability committee (DAC)” (which guarantees off-chain data is not lost) — and proposes secure designs for each. It uses formal verification to address the sequencer’s single-point-of-failure problem and the DAC’s collusion risk (where multiple members conspire to destroy data). This is foundational research directly relevant to Cardano’s Hydra and future rollup integration security design.
#20 Towards a Formal Foundation for Blockchain ZK Rollups
📍 2025 ／ 👤 IOG authors ／ #rollups #formalmethods #zk
This paper formally analyzes the safety, censorship resistance, and upgradeability of ZK rollups (Layer 2 technology that uses zero-knowledge proofs to validate off-chain transactions) using the Alloy model checking language. The paper enumerates design flaws in existing ZK rollup implementations (such as Ethereum’s zkSync) and identifies multisig attack vectors that can result in total user fund loss, as well as forced transaction queue inconsistencies. An improved model that addresses these issues has been verified through model checking. The paper also serves as a practical guide for anticipating similar pitfalls in Cardano’s ZK infrastructure design.
Hydra & Interchains
#21 Interhead Hydra: Two Heads are Better than One
📍 December 2025 ／ 👤 Badertscher, Coretti, Gaži, Kiayias, Russell ／ #hydra #state-channels
This paper proposes the “Interhead” construction, which extends Cardano’s Hydra scaling solution to support cross-Head transactions. Traditional Hydra Heads are independent multi-party state channels, and executing transactions between different Heads required routing through Layer 1. Interhead resolves this limitation by iteratively creating “virtual Hydra Heads” between existing Heads, allowing any number of parties to efficiently conduct cross-Head transactions while sharing collateral costs across multiple intermediaries. It is the first construction in the history of multi-party state channels to support channels with an arbitrary number of parties.
Formal Methods & Programming Languages
#22 A Layered Certifying Compiler Architecture
📍 FUNARCH 2025 ／ 👤 Krijnen, Swierstra, Chakravarty, Dral, Keller ／ #smartcontract #formalmethods
This paper proposes a layered architecture for guaranteeing the correctness of smart contract compilers — specifically Cardano’s Plinth compiler — through “translation validation.” Classical verification approaches like CompCert or CakeML require implementing the entire compiler in a proof assistant, which is impractical for actively maintained open-source projects. Instead, this paper proposes developing an independent “certifier” that post-hoc verifies the correctness of each compilation. This approach can reduce the risk of smart contract vulnerabilities caused by Plinth compiler bugs.
#23 Plinth: A Plugin-Powered Language Built on Haskell (Experience Report)
📍 Haskell 2025 ／ 👤 Liu, MacKenzie, Kireev, Peyton Jones, Wadler, Chakravarty ／ #plutus #haskell
An experience report on designing and implementing Plinth — the on-chain compiler for Cardano smart contracts — as a GHC plugin. Plinth (formerly known as PlutusTx) compiles Haskell code to Cardano’s Plutus Core (a functional language based on Fωμ type theory). The paper reports on the benefits and challenges of the GHC plugin architecture — including integration with type inference, intervention in optimization passes, and improving error messages — based on actual development experience. This is practical research that directly improves the developer experience for smart contract developers.
#24 Automated Verification of Proofs in the Universal Composability Framework with Markov Decision Processes
📍 CANS 2025 ／ 👤 Jourenko, Völker ／ #universal-composition #formalmethods
This paper proposes a method for automating parts of the proof process in the Universal Composability (UC) framework — widely used for security proofs of cryptographic protocols — using Markov Decision Processes (MDPs). UC framework proofs are complex, often requiring manual verification and thus prone to errors. The paper builds a system that uses MDPs to automatically generate and verify parts of UC proofs, aiming to improve both proof reliability and reduce manual effort. Since many IOG-built protocols — including Ouroboros, Hydra, and governance systems — have their security discussed within the UC framework, automated verification tooling directly benefits both research productivity and overall security.
Summary by Categories
| Theme | Count | Key Topics |
|---|---|---|
| Consensus & Ouroboros | 4 | Leios, Peras, Grinding Attacks, Proof-of-Useful-Work |
| Fairness & Fee Design | 3 | Transaction Order Fairness, Multi-dimensional Pricing, Tiered Fees |
| Game Theory & Mechanism Design | 7 | DRep Rewards, Stake Pools, Tokenomics, Geographic Diversity, Participatory Budgeting, LLM Preference Elicitation |
| Zero-Knowledge Proofs & Cryptography | 2 | UC-Secure SNARKs, Crosschain ZK Proofs |
| Networking & Distributed Systems | 2 | P2P Overlay Optimization, Setchain Scalability |
| Rollups & Layer 2 | 2 | Secure Sequencer & DAC, ZK Rollup Formal Foundation |
| Hydra & Interchains | 1 | Interhead Construction, Cross-Head Transactions |
| Formal Methods & Programming Languages | 3 | Certifying Compiler, Plinth, UC Proof Automation |
※ Some papers were published on arXiv or similar preprint servers in 2024 and formally accepted at conferences in 2025. They are counted as 2025 papers based on IOG’s official library listing criteria. For links to the full papers, please visit iog.io/papers.
本稿はカルダノステークプール「CoffeePool☕️」「CardanoKissa☕️」が作成しました。
COFFEの活動を応援いただける方はぜひ、委任（Delegate）のご検討をお願いいたします😊
🗳️「DRep ID」もぜひご検討ください！（クリックでコピーできます↓）
CoffeePoolならびにKISSAは、5年以上の安定稼働と1万ブロック以上の生成実績を持つステークプールです。
初心者でも長期ホルダーでも、楽しくカルダノ について語り合える discordスペースを作りました😊☕️
・なりすましやscam対策のためXアカウント認証で運営😊
・初心者ホルダーさん大歓迎！☕️
SPOも続々参加しています！ ぜひお気軽にお立ち寄りください。
参加方法は☟
discord.gg/TNy7QNua7c
