arXiv subjects

Roger Wattenhofer

Publications and source records attributed to Roger Wattenhofer.

At least 253 records · Page 14Linked to original sources

arXiv2018

Structure and Content of the Visible Darknet

In this paper, we analyze the topology and the content found on the "darknet", the set of websites accessible via Tor. We created a darknet spider and crawled the darknet starting from a bootstrap list by recursively following links. We explored the whole connected component of more than 34,000 hidden services, of which we found 10,000 to be online. Contrary to folklore belief, the visible part of the darknet is surprisingly well-connected through hub websites such as wikis and forums. We performed a comprehensive categorization of the content using supervised machine learning. We observe that about half of the visible dark web content is related to apparently licit activities based on our classifier. A significant amount of content pertains to software repositories, blogs, and activism-related websites. Among unlawful hidden services, most pertain to fraudulent websites, services selling counterfeit goods, and drug markets.

cs.CY↗

arXiv2018

High Dimensional Clustering with $r$-nets

Clustering, a fundamental task in data science and machine learning, groups a set of objects in such a way that objects in the same cluster are closer to each other than to those in other clusters. In this paper, we consider a well-known structure, so-called $r$-nets, which rigorously captures the properties of clustering. We devise algorithms that improve the run-time of approximating $r$-nets in high-dimensional spaces with $\ell_1$ and $\ell_2$ metrics from $\tilde{O}(dn^{2-Θ(\sqrtε)})$ to $\tilde{O}(dn + n^{2-α})$, where $α= Ω({ε^{1/3}}/{\log(1/ε)})$. These algorithms are also used to improve a framework that provides approximate solutions to other high dimensional distance problems. Using this framework, several important related problems can also be solved efficiently, e.g., $(1+ε)$-approximate $k$th-nearest neighbor distance, $(4+ε)$-approximate Min-Max clustering, $(4+ε)$-approximate $k$-center clustering. In addition, we build an algorithm that $(1+ε)$-approximates greedy permutations in time $\tilde{O}((dn + n^{2-α}) \cdot \logΦ)$ where $Φ$ is the spread of the input. This algorithm is used to $(2+ε)$-approximate $k$-center with the same time complexity.

cs.CG↗

arXiv2018

Payment Network Design with Fees

Payment channels are the most prominent solution to the blockchain scalability problem. We introduce the problem of network design with fees for payment channels from the perspective of a Payment Service Provider (PSP). Given a set of transactions, we examine the optimal graph structure and fee assignment to maximize the PSP's profit. A customer prefers to route transactions through the PSP's network if the cheapest path from sender to receiver is financially interesting, i.e., if the path costs less than the blockchain fee. When the graph structure is a tree, and the PSP facilitates all transactions, the problem can be formulated as a linear program. For a path graph, we present a polynomial time algorithm to assign optimal fees. We also show that the star network, where the center is an additional node acting as an intermediary, is a near-optimal solution to the network design problem.

cs.DS↗

arXiv2018

Algorithmic Blockchain Channel Design

Payment networks, also known as channels, are a most promising solution to the throughput problem of cryptocurrencies. In this paper we study the design of capital-efficient payment networks, offline as well as online variants. We want to know how to compute an efficient payment network topology, how capital should be assigned to the individual edges, and how to decide which transactions to accept. Towards this end, we present a flurry of interesting results, basic but generally applicable insights on the one hand, and hardness results and approximation algorithms on the other hand.

cs.DS↗

arXiv2018

Using State Predictions for Value Regularization in Curiosity Driven Deep Reinforcement Learning

Learning in sparse reward settings remains a challenge in Reinforcement Learning, which is often addressed by using intrinsic rewards. One promising strategy is inspired by human curiosity, requiring the agent to learn to predict the future. In this paper a curiosity-driven agent is extended to use these predictions directly for training. To achieve this, the agent predicts the value function of the next state at any point in time. Subsequently, the consistency of this prediction with the current value function is measured, which is then used as a regularization term in the loss function of the algorithm. Experiments were made on grid-world environments as well as on a 3D navigation task, both with sparse rewards. In the first case the extended agent is able to learn significantly faster than the baselines.

cs.LG↗

arXiv2018

The Urban Last Mile Problem: Autonomous Drone Delivery to Your Balcony

Drone delivery has been a hot topic in the industry in the past few years. However, existing approaches either focus on rural areas or rely on centralized drop-off locations from where the last mile delivery is performed. In this paper we tackle the problem of autonomous last mile delivery in urban environments using an off-the-shelf drone. We build a prototype system that is able to fly to the approximate delivery location using GPS and then find the exact drop-off location using visual navigation. The drop-off location could, e.g., be on a balcony or porch, and simply needs to be indicated by a visual marker on the wall or window. We test our system components in simulated environments, including the visual navigation and collision avoidance. Finally, we deploy our drone in a real-world environment and show how it can find the drop-off point on a balcony. To stimulate future research in this topic we open source our code.

cs.RO↗

arXiv2018

Symbolic Music Genre Transfer with CycleGAN

Deep generative models such as Variational Autoencoders (VAEs) and Generative Adversarial Networks (GANs) have recently been applied to style and domain transfer for images, and in the case of VAEs, music. GAN-based models employing several generators and some form of cycle consistency loss have been among the most successful for image domain transfer. In this paper we apply such a model to symbolic music and show the feasibility of our approach for music genre transfer. Evaluations using separate genre classifiers show that the style transfer works well. In order to improve the fidelity of the transformed music, we add additional discriminators that cause the generators to keep the structure of the original music mostly intact, while still achieving strong genre transfer. Visual and audible results further show the potential of our approach. To the best of our knowledge, this paper represents the first application of GANs to symbolic music domain transfer.

cs.SD↗

arXiv2018

MIDI-VAE: Modeling Dynamics and Instrumentation of Music with Applications to Style Transfer

We introduce MIDI-VAE, a neural network model based on Variational Autoencoders that is capable of handling polyphonic music with multiple instrument tracks, as well as modeling the dynamics of music by incorporating note durations and velocities. We show that MIDI-VAE can perform style transfer on symbolic music by automatically changing pitches, dynamics and instruments of a music piece from, e.g., a Classical to a Jazz style. We evaluate the efficacy of the style transfer by training separate style validation classifiers. Our model can also interpolate between short pieces of music, produce medleys and create mixtures of entire songs. The interpolations smoothly change pitches, dynamics and instrumentation to create a harmonic bridge between two music pieces. To the best of our knowledge, this work represents the first successful attempt at applying neural style transfer to complete musical compositions.

cs.SD↗

arXiv2018

Reducing Compare-and-Swap to Consensus Number One Primitives

The consensus number of an object is the maximum number of processes among which binary consensus can be solved using any number of instances of the object and read-write registers. Herlihy [6] showed in his seminal work that if an object has a consensus number of n, then there is a universal construction for a wait-free and linearizable implementation of any non-trivial concurrent object or data structure that is shared among n processes. Thus, a synchronization object such as compare-and-swap with an infinite consensus number and the corresponding instruction can be viewed as "strong". On the other hand, a synchronization object such as fetch-and-add with consensus number two and the corresponding fetch-and-add instruction can be viewed as "weak". Ellen et al. [2] observed recently that an object supporting two weak instructions can also achieve infinite consensus number like an object that supports one strong instruction. Using Herlihy's universal construction, this implies that ignoring concerns about efficiency, one can design any concurrent data structure or algorithm using only weak instructions. However, is it possible that a combination of weak instructions is really powerful enough to efficiently replace a strong instruction, like compare-and-swap, without incurring a large overhead in time or space? In this paper, we answer this question by giving an O(1) time wait-free and linearizable implementation of a compare-and-swap register shared among n processes using read-write registers and O(1) registers that support two synchronization primitives half-max and max-write, each having consensus number one. Thus, any algorithm that solves some arbitrary synchronization problem using read-write and compare-and-swap registers can be transformed into an algorithm that has the same asymptotic time complexity and only uses consensus number one instructions.

cs.DS↗

arXiv2018

Tight Bounds for Asynchronous Collaborative Grid Exploration

Consider a small group of mobile agents whose goal is to locate a certain cell in a two-dimensional infinite grid. The agents operate in an asynchronous environment, where in each discrete time step, an arbitrary subset of the agents execute one atomic look-compute-move cycle. The protocol controlling each agent is determined by a (possibly distinct) finite automaton. The only means of communication is to sense the states of the agents sharing the same grid cell. Whenever an agent moves, the destination cell of the movement is chosen by the agent's automaton from the set of neighboring grid cells. We study the minimum number of agents required to locate the target cell within finite time and our main result states a tight lower bound for agents endowed with a global compass. Furthermore, we show that the lack of such a compass makes the problem strictly more difficult and present tight upper and lower bounds for this case.

cs.DC↗

arXiv2018

Byzantine Preferential Voting

In the Byzantine agreement problem, n nodes with possibly different input values aim to reach agreement on a common value in the presence of t < n/3 Byzantine nodes which represent arbitrary failures in the system. This paper introduces a generalization of Byzantine agreement, where the input values of the nodes are preference rankings of three or more candidates. We show that consensus on preferences, which is an important question in social choice theory, complements already known results from Byzantine agreement. In addition preferential voting raises new questions about how to approximate consensus vectors. We propose a deterministic algorithm to solve Byzantine agreement on rankings under a generalized validity condition, which we call Pareto-Validity. These results are then extended by considering a special voting rule which chooses the Kemeny median as the consensus vector. For this rule, we derive a lower bound on the approximation ratio of the Kemeny median that can be guaranteed by any deterministic algorithm. We then provide an algorithm matching this lower bound. To our knowledge, this is the first non-trivial multi-dimensional approach which can tolerate a constant fraction of Byzantine nodes.

cs.DC↗

arXiv2018

Natural Language Multitasking: Analyzing and Improving Syntactic Saliency of Hidden Representations

We train multi-task autoencoders on linguistic tasks and analyze the learned hidden sentence representations. The representations change significantly when translation and part-of-speech decoders are added. The more decoders a model employs, the better it clusters sentences according to their syntactic similarity, as the representation space becomes less entangled. We explore the structure of the representation space by interpolating between sentences, which yields interesting pseudo-English sentences, many of which have recognizable syntactic structure. Lastly, we point out an interesting property of our models: The difference-vector between two sentences can be added to change a third sentence with similar features in a meaningful way.

cs.CL↗

arXiv2017

JamBot: Music Theory Aware Chord Based Generation of Polyphonic Music with LSTMs

We propose a novel approach for the generation of polyphonic music based on LSTMs. We generate music in two steps. First, a chord LSTM predicts a chord progression based on a chord embedding. A second LSTM then generates polyphonic music from the predicted chord progression. The generated music sounds pleasing and harmonic, with only few dissonant notes. It has clear long-term structure that is similar to what a musician would play during a jam session. We show that our approach is sensible from a music theory perspective by evaluating the learned chord embeddings. Surprisingly, our simple model managed to extract the circle of fifths, an important tool in music theory, from the dataset.

cs.SD↗

arXiv2017

Teaching a Machine to Read Maps with Deep Reinforcement Learning

The ability to use a 2D map to navigate a complex 3D environment is quite remarkable, and even difficult for many humans. Localization and navigation is also an important problem in domains such as robotics, and has recently become a focus of the deep reinforcement learning community. In this paper we teach a reinforcement learning agent to read a map in order to find the shortest way out of a random maze it has never seen before. Our system combines several state-of-the-art methods such as A3C and incorporates novel elements such as a recurrent localization cell. Our agent learns to localize itself based on 3D first person images and an approximate orientation angle. The agent generalizes well to bigger mazes, showing that it learned useful localization and navigation capabilities.

cs.RO↗

arXiv2017

Towards Reduced Instruction Sets for Synchronization

Contrary to common belief, a recent work by Ellen, Gelashvili, Shavit, and Zhu has shown that computability does not require multicore architectures to support "strong" synchronization instructions like compare-and-swap, as opposed to combinations of "weaker" instructions like decrement and multiply. However, this is the status quo, and in turn, most efficient concurrent data-structures heavily rely on compare-and-swap (e.g. for swinging pointers and in general, conflict resolution). We show that this need not be the case, by designing and implementing a concurrent linearizable Log data-structure (also known as a History object), supporting two operations: append(item), which appends the item to the log, and get-log(), which returns the appended items so far, in order. Readers are wait-free and writers are lock-free, and this data-structure can be used in a lock-free universal construction to implement any concurrent object with a given sequential specification. Our implementation uses atomic read, xor, decrement, and fetch-and-increment instructions supported on X86 architectures, and provides similar performance to a compare-and-swap-based solution on today's hardware. This raises a fundamental question about minimal set of synchronization instructions that the architectures have to support.

cs.DC↗

arXiv2016

Opening the Frey/Osborne Black Box: Which Tasks of a Job are Susceptible to Computerization?

In their seminal paper, Frey and Osborne quantified the automation of jobs, by assigning each job in the O*NET database a probability to be automated. In this paper, we refine their results in the following way: Every O*NET job consists of a set of tasks, and these tasks can be related. We use a linear program to assign probabilities to tasks, such that related tasks have a similar probability and the tasks can explain the computerization probability of a job. Analyzing jobs on the level of tasks helps comprehending the results, as experts as well as laymen can more easily criticize and refine what parts of a job are susceptible to computerization.

cs.CY↗

arXiv2016

CLEX: Yet Another Supercomputer Architecture?

We propose the CLEX supercomputer topology and routing scheme. We prove that CLEX can utilize a constant fraction of the total bandwidth for point-to-point communication, at delays proportional to the sum of the number of intermediate hops and the maximum physical distance between any two nodes. Moreover, % applying an asymmetric bandwidth assignment to the links, all-to-all communication can be realized $(1+o(1))$-optimally both with regard to bandwidth and delays. This is achieved at node degrees of $n^{\varepsilon}$, for an arbitrary small constant $\varepsilon\in (0,1]$. In contrast, these results are impossible in any network featuring constant or polylogarithmic node degrees. Through simulation, we assess the benefits of an implementation of the proposed communication strategy. Our results indicate that, for a million processors, CLEX can increase bandwidth utilization and reduce average routing path length by at least factors $10$ respectively $5$ in comparison to a torus network. Furthermore, the CLEX communication scheme features several other properties, such as deadlock-freedom, inherent fault-tolerance, and canonical partition into smaller subsystems.

cs.DC↗

arXiv2016

Distributed Local Multi-Aggregation and Centrality Approximation

We study local aggregation and graph analysis in distributed environments using the message passing model. We provide a flexible framework, where each of the nodes in a set $S$--which is a subset of all nodes in the network--can perform a large range of common aggregation functions in its $k$-neighborhood. We study this problem in the CONGEST model, where in each synchronous round, every node can transmit a different (but short) message to each of its neighbors. While the $k$-neighborhoods of nodes in $S$ might overlap and aggregation could cause congestion in this model, we present an algorithm that needs time $O(|S|+k)$ even when each of the nodes in $S$ performs a different aggregation on its $k$-neighborhood. The framework is not restricted to aggregation-trees such that it can be used for more advanced graph analysis. We demonstrate this by providing efficient approximations of centrality measures and approximation of minimum routing cost trees.

cs.DC↗