arXiv subjects

Anna Huber

Publications and source records attributed to Anna Huber.

7 recordsLinked to original sources

arXiv2026

Energy-Based Injury Protection Database: Including Shearing Contact Thresholds for Hand and Finger Using Porcine Surrogates

While robotics research continues to propose strategies for collision avoidance in human-robot interaction, the reality of constrained environments and future humanoid systems makes contact inevitable. To mitigate injury risks, energy-constraining control approaches are commonly used, often relying on safety thresholds derived from blunt impact data in EN ISO 10218-2:2025. However, this dataset does not extend to edged or pointed collisions. Without scalable, clinically grounded datasets covering diverse contact scenarios, safety validation remains limited. Previous studies have laid the groundwork by assessing surrogate-based velocity and mass limits across various geometries, focusing on perpendicular impacts. This study expands those datasets by including shearing contact scenarios in unconstrained collisions, revealing that collision angle significantly affects injury outcomes. Notably, unconstrained shearing contacts result in fewer injuries than perpendicular ones. By reevaluating all prior porcine surrogate data, we establish energy thresholds across geometries and contact types, forming the first energy-based Injury Protection Database. This enables the development of meaningful energy-limiting controllers that ensure safety across a wide range of realistic collision events.

cs.RO↗

arXiv2025

A Women's Health Benchmark for Large Language Models

As large language models (LLMs) become primary sources of health information for millions, their accuracy in women's health remains critically unexamined. We introduce the Women's Health Benchmark (WHB), the first benchmark evaluating LLM performance specifically in women's health. Our benchmark comprises 96 rigorously validated model stumps covering five medical specialties (obstetrics and gynecology, emergency medicine, primary care, oncology, and neurology), three query types (patient query, clinician query, and evidence/policy query), and eight error types (dosage/medication errors, missing critical information, outdated guidelines/treatment recommendations, incorrect treatment advice, incorrect factual information, missing/incorrect differential diagnosis, missed urgency, and inappropriate recommendations). We evaluated 13 state-of-the-art LLMs and revealed alarming gaps: current models show approximately 60\% failure rates on the women's health benchmark, with performance varying dramatically across specialties and error types. Notably, models universally struggle with "missed urgency" indicators, while newer models like GPT-5 show significant improvements in avoiding inappropriate recommendations. Our findings underscore that AI chatbots are not yet fully able of providing reliable advice in women's health.

cs.CL↗

arXiv2021

Efficient and Accurate In-Database Machine Learning with SQL Code Generation in Python

Following an analysis of the advantages of SQL-based Machine Learning (ML) and a short literature survey of the field, we describe a novel method for In-Database Machine Learning (IDBML). We contribute a process for SQL-code generation in Python using template macros in Jinja2 as well as the prototype implementation of the process. We describe our implementation of the process to compute multidimensional histogram (MDH) probability estimation in SQL. For this, we contribute and implement a novel discretization method called equal quantized rank binning (EQRB) and equal-width binning (EWB). Based on this, we provide data gathered in a benchmarking experiment for the quantitative empirical evaluation of our method and system using the Covertype dataset. We measured accuracy and computation time and compared it to Scikit Learn state of the art classification algorithms. Using EWB, our multidimensional probability estimation was the fastest of all tested algorithms, while being only 1-2% less accurate than the best state of the art methods found (decision trees and random forests). Our method was significantly more accurate than Naive Bayes, which assumes independent one-dimensional probabilities and/or densities. Also, our method was significantly more accurate and faster than logistic regression. This motivates for further research in accuracy improvement and in IDBML with SQL code generation for big data and larger-than-memory datasets.

cs.DB↗

arXiv2013

Towards Minimizing k-Submodular Functions

In this paper we investigate k-submodular functions. This natural family of discrete functions includes submodular and bisubmodular functions as the special cases k = 1 and k = 2 respectively. In particular we generalize the known Min-Max-Theorem for submodular and bisubmodular functions. This theorem asserts that the minimum of the (bi)submodular function can be found by solving a maximization problem over a (bi)submodular polyhedron. We define and investigate a k-submodular polyhedron and prove a Min-Max-Theorem for k-submodular functions.

cs.DM↗

arXiv2013

Oracle Tractability of Skew Bisubmodular Functions

In this paper we consider skew bisubmodular functions as introduced in [9]. We construct a convex extension of a skew bisubmodular function which we call Lov\'asz extension in correspondence to the submodular case. We use this extension to show that skew bisubmodular functions given by an oracle can be minimised in polynomial time.

cs.CC↗

arXiv2010

Strong Robustness of Randomized Rumor Spreading Protocols

Randomized rumor spreading is a classical protocol to disseminate information across a network. At SODA 2008, a quasirandom version of this protocol was proposed and competitive bounds for its run-time were proven. This prompts the question: to what extent does the quasirandom protocol inherit the second principal advantage of randomized rumor spreading, namely robustness against transmission failures? In this paper, we present a result precise up to $(1 \pm o(1))$ factors. We limit ourselves to the network in which every two vertices are connected by a direct link. Run-times accurate to their leading constants are unknown for all other non-trivial networks. We show that if each transmission reaches its destination with a probability of $p \in (0,1]$, after $(1+\e)(\frac{1}{\log_2(1+p)}\log_2n+\frac{1}{p}\ln n)$ rounds the quasirandom protocol has informed all $n$ nodes in the network with probability at least $1-n^{-p\e/40}$. Note that this is faster than the intuitively natural $1/p$ factor increase over the run-time of approximately $\log_2 n + \ln n $ for the non-corrupted case. We also provide a corresponding lower bound for the classical model. This demonstrates that the quasirandom model is at least as robust as the fully random model despite the greatly reduced degree of independent randomness.

cs.DM↗

arXiv2009

The Speed of Broadcasting in Random Networks: Density Does Not Matter

Broadcasting algorithms are of fundamental importance for distributed systems engineering. In this paper we revisit the classical and well-studied push protocol for message broadcasting. Assuming that initially only one node has some piece of information, at each stage every one of the informed nodes chooses randomly and independently one of its neighbors and passes the message to it. The performance of the push protocol on a fully connected network, where each node is joined by a link to every other node, is very well understood. In particular, Frieze and Grimmett proved that with probability 1-o(1) the push protocol completes the broadcasting of the message within (1 +/- ε) (log_2 n + ln n) stages, where n is the number of nodes of the network. However, there are no tight bounds for the broadcast time on networks that are significantly sparser than the complete graph. In this work we consider random networks on n nodes, where every edge is present with probability p, independently of every other edge. We show that if p > f(n)ln n/ n, where f(n) is any function that tends to infinity as n grows, then the push protocol broadcasts the message within (1 +/- ε) (log_2 n + ln n) stages with probability 1-o(1). In other words, in almost every network of density d such that d > f(n)ln n, the push protocol broadcasts a message as fast as in a fully connected network. This is quite surprising in the sense that the time needed remains essentially unaffected by the fact that most of the links are missing.

math.CO↗