# About Me:

–I am a second-year PhD student in the Applied Crypto Group and am advised by Dan Boneh

–My main reserach interests are in cryptography, game theory and cryptocurrencies.

–I am an avid runner and train and compete for Strava Track Club

# Publications (Google Scholar):

## Cryptography and Cryptocurrencies

### Bulletproofs: Short Proofs for Confidential Transactions and More

Authors: B. Bünz, J. Bootle, D. Boneh, Andrew Poelstra, Pieter Wuille and Greg Maxwell

Published at Oakland 2018

libsecp256k1 implementation by Andrew Poelstra

TLDR: Confidential transactions are Bitcoin transactions which are publicly verifiable but do not reveal the amounts that are transferred. They rely on cryptographic commitments and so called zero-knowledge proofs of knowledge. We present a new kind of zero-knowledge proof which is much more efficient and can be used to drastically reduce the size of confidential transactions. On a more technical note bulletproofs are non-interactive zero knowledge proofs without trusted setup and with only logarithmic proof size. Proving and verification cost are linear with low constant overhead.

### Provisions: Privacy-preserving proofs of solvency for Bitcoin exchanges

Authors: G. Dagher, B. Bünz, J.Bonneau, J.Clark and D. Boneh

Published at CCS 2015

Talk at Next Context Conference

TLDR: How can a Bitcoin exchange proof that they have enough funds to satisfy all their customers demands without revealing the customers balances, the bitcoin addresses they control or even the total amount of bitcoin they have.

### Proofs-of-delay and randomness beacons in Ethereum

Authors: B. Bünz, S. Goldfeder, J. Bonneau

Presented at IEEE S&B Workshop

TLDR: We show how one can generate an unpredictable randomness beacon that is publicly verifiable using a blockchain. The beacon can be used to verify the correct execution of randomized algorithms such as lotteries. The novel property of the beacon is that it is publicly verifiable in that a verifier is convinced that the beacon was unpredictable even if she did not partake in the generation of the beacon and without any trust assumptions. We also show how we can enable interactive verification using an efficient smart contract.

## Game Theory (Combinatorial Auctions)

### Designing Core-Selecting Payment Rules: A Computational Search Approach

Authors: B. Bünz, B. Lubin, S. Seuken

Published at EC 2018

### A Faster Core Constraint Generation Algorithm for Combinatorial Auctions

Authors: B. Bünz, B. Lubin, S. Seuken

Published at AAAI 2015

TLDR: We significantly improve on the current state of the art algorithm for computing combinatorial auctions. These auctions were multiple related goods are sold in the same auction are for example used to allocated spectrum to cellular companies around the world. These auctions often generate billions of dollars in revenue but are often limited to a small number of bidders and goods. Faster algorithms for computing their outcome will enable larger scale applications.

### Computing Bayes-Nash Equilibria in Combinatorial Auctions with Continuous Value and Action Spaces

Authors: V. Bosshard, B. Bünz, B. Lubin, S. Seuken

Published at IJCAI 2017

### LLG BNEs (Working Paper)

Authors: B. Lubin, B. Bünz, S. Seuken

Extended abstract published at AMMA 2015

## Artificial Intelligence

### Learning a SAT Solver from Single-Bit Supervision

Authors: D. Selsam, M. Lamm, B. Bünz, P. Liang, L. de Moura,D. Dill

Talk by Daniel Selsam at Microsoft Research

TLDR: We develop a neural network based solver for finding satisfying assingments to boolean formulas (SAT solver). At training time the network is given satisfying formulas and only the information of whether the formula has a solution or not. Despite this minimal supervision we are able to directly read of satisftying assignments from the activations of the network if it classifies a formula as satisfiable. Additionally we can even find contradictions if the formula is unsatisfiable. Given that classifying boolean formulas is an NP-complete problem this an interesting exploration into the abilities and flexibilities of neural network and also raises interesting possibilities of using neural networks in the development of state of the art SAT solvers.