Multiple sclerosis (MS) is a debilitating neurological disease affecting nearly one million people in the United States. Sphingosine-1-phosphate receptor 1, or S1PR1, is a protein target for MS. Siponimod, a ligand of S1PR1, was approved by the FDA in 2019 for MS treatment, but there is a demonstrated need for better therapies. To this end, we finetuned an autoencoder machine learning model that converts chemical formulas into mathematical vectors and generated over 500 molecular variants based on siponimod, out of which 25 compounds had higher predicted binding affinity to S1PR1. The model was able to generate these ligands in just under one hour. Filtering these compounds led to the discovery of six promising candidates with good drug-like properties and ease of synthesis. Furthermore, by analyzing the binding interactions for these ligands, we uncovered several chemical properties that contribute to high binding affinity to S1PR1. This study demonstrates that machine learning can accelerate the drug discovery process and reveal new insights into protein-drug interactions.

As machine learning (ML) systems increasingly permeate high-stakes settings such as healthcare, transportation, military, and national security, concerns regarding their reliability have emerged. Despite notable progress, the performance of these systems can significantly diminish due to adversarial attacks or environmental changes, leading to overconfident predictions, failures to detect input faults, and an inability to generalize in unexpected scenarios. This paper proposes a holistic assessment methodology for the reliability of ML systems. Our framework evaluates five key properties: in-distribution accuracy, distribution-shift robustness, adversarial robustness, calibration, and out-of-distribution detection. A reliability score is also introduced and used to assess the overall system reliability. To provide insights into the performance of different algorithmic approaches, we identify and categorize state-of-the-art techniques, then evaluate a selection on real-world tasks using our proposed reliability metrics and reliability score. Our analysis of over 500 models reveals that designing for one metric does not necessarily constrain others but certain algorithmic techniques can improve reliability across multiple metrics simultaneously. This study contributes to a more comprehensive understanding of ML reliability and provides a roadmap for future research and development.

This paper presents two self-contained tutorials on stance detection in Twitter data using BERT fine-tuning and prompting large language models (LLMs). The first tutorial explains BERT architecture and tokenization, guiding users through training, tuning, and evaluating standard and domain-specific BERT models with HuggingFace transformers. The second focuses on constructing prompts and few-shot examples to elicit stances from ChatGPT and open-source FLAN-T5 without fine-tuning. Various prompting strategies are implemented and evaluated using confusion matrices and macro F1 scores. The tutorials provide code, visualizations, and insights revealing the strengths of few-shot ChatGPT and FLAN-T5 which outperform fine-tuned BERTs. By covering both model fine-tuning and prompting-based techniques in an accessible, hands-on manner, these tutorials enable learners to gain applied experience with cutting-edge methods for stance detection.

Figure 1: We explore how autonomous robot adjustments within task tolerances shape task performance and user experience in mimicry-based telemanipulation. For example, (A) in a teleoperated writing task, our human-subject experiment results indicate that (B) if the robot autonomously tilts or rotates the pen within task tolerances to improve pen tip accuracy and motion smoothness, telemanipulation is enhanced by the high-quality robot motions enabled by task tolerances, despite users lacking full control of the robot. Abstract-- We explore task tolerances, i.e., allowable position In this paper, we explore task tolerances as an important resource to facilitate functional mimicry in telemanipulation. The functional mimicry paradigm allows a robot to autonomously adjust within tolerances to generate more Mimicry-based telemanipulation maps a human operator's accurate, smooth, and feasible motions. In our previous hand movement to a robot's end effector in real time [1], work, we presented RangedIK [4] as a real-time motion [2]. The robot is often required to mimic the operator's generation method that exploits flexibility afforded by task movement as exactly as possible, so we call this paradigm tolerances.

Most group fairness notions detect unethical biases by computing statistical parity metrics on a model's output. However, this approach suffers from several shortcomings, such as philosophical disagreement, mutual incompatibility, and lack of interpretability. These shortcomings have spurred the research on complementary bias detection methods that offer additional transparency into the sources of discrimination and are agnostic towards an a priori decision on the definition of fairness and choice of protected features. A recent proposal in this direction is LUCID (Locating Unfairness through Canonical Inverse Design), where canonical sets are generated by performing gradient descent on the input space, revealing a model's desired input given a preferred output. This information about the model's mechanisms, i.e., which feature values are essential to obtain specific outputs, allows exposing potential unethical biases in its internal logic. Here, we present LUCID-GAN, which generates canonical inputs via a conditional generative model instead of gradient-based inverse design. LUCID-GAN has several benefits, including that it applies to non-differentiable models, ensures that canonical sets consist of realistic inputs, and allows to assess proxy and intersectional discrimination. We empirically evaluate LUCID-GAN on the UCI Adult and COMPAS data sets and show that it allows for detecting unethical biases in black-box models without requiring access to the training data.

We introduce AbDiffuser, an equivariant and physics-informed diffusion model for the joint generation of antibody 3D structures and sequences. AbDiffuser is built on top of a new representation of protein structure, relies on a novel architecture for aligned proteins, and utilizes strong diffusion priors to improve the denoising process. Our approach improves protein diffusion by taking advantage of domain knowledge and physics-based constraints; handles sequence-length changes; and reduces memory complexity by an order of magnitude enabling backbone and side chain generation. Numerical experiments showcase the ability of AbDiffuser to generate antibodies that closely track the sequence and structural properties of a reference set. Laboratory experiments confirm that all 16 HER2 antibodies discovered were expressed at high levels and that 57.1% of selected designs were tight binders. We focus on the generation of immunoglobulin proteins, also known as antibodies, that help the immune ...

The IceCube South Pole Neutrino Observatory is a Cherenkov detector instrumented in a cubic kilometer of ice at the South Pole. IceCube's primary scientific goal is the detection of TeV neutrino emissions from astrophysical sources. At the lower center of the IceCube array, there is a subdetector called DeepCore, which has a denser configuration that makes it possible to lower the energy threshold of IceCube and observe GeV-scale neutrinos, opening the window to atmospheric neutrino oscillations studies. Advances in physics sensitivity have recently been achieved by employing Convolutional Neural Networks to reconstruct neutrino interactions in the DeepCore detector. In this contribution, the recent IceCube result from the atmospheric muon neutrino disappearance analysis using the CNN-reconstructed neutrino sample is presented and compared to the existing worldwide measurements.

Recent advances of powerful Language Models have allowed Natural Language Generation (NLG) to emerge as an important technology that can not only perform traditional tasks like summarisation or translation, but also serve as a natural language interface to a variety of applications. As such, it is crucial that NLG systems are trustworthy and reliable, for example by indicating when they are likely to be wrong; and supporting multiple views, backgrounds and writing styles -- reflecting diverse human sub-populations. In this paper, we argue that a principled treatment of uncertainty can assist in creating systems and evaluation protocols better aligned with these goals. We first present the fundamental theory, frameworks and vocabulary required to represent uncertainty. We then characterise the main sources of uncertainty in NLG from a linguistic perspective, and propose a two-dimensional taxonomy that is more informative and faithful than the popular aleatoric/epistemic dichotomy. Finally, we move from theory to applications and highlight exciting research directions that exploit uncertainty to power decoding, controllable generation, self-assessment, selective answering, active learning and more.

Human interactions create social networks forming the backbone of societies. Individuals adjust their opinions by exchanging information through social interactions. Two recurrent questions are whether social structures promote opinion polarisation or consensus in societies and whether polarisation can be avoided, particularly on social media. In this paper, we hypothesise that not only network structure but also the timings of social interactions regulate the emergence of opinion clusters. We devise a temporal version of the Deffuant opinion model where pairwise interactions follow temporal patterns and show that burstiness alone is sufficient to refrain from consensus and polarisation by promoting the reinforcement of local opinions. Individuals self-organise into a multi-partisan society due to network clustering, but the diversity of opinion clusters further increases with burstiness, particularly when individuals have low tolerance and prefer to adjust to similar peers. The emergent opinion landscape is well-balanced regarding clusters' size, with a small fraction of individuals converging to extreme opinions. We thus argue that polarisation is more likely to emerge in social media than offline social networks because of the relatively low social clustering observed online. Counter-intuitively, strengthening online social networks by increasing social redundancy may be a venue to reduce polarisation and promote opinion diversity.

Even though machine learning (ML) pipelines affect an increasing array of stakeholders, there is little work on how input from stakeholders is recorded and incorporated. We propose FeedbackLogs, addenda to existing documentation of ML pipelines, to track the input of multiple stakeholders. Each log records important details about the feedback collection process, the feedback itself, and how the feedback is used to update the ML pipeline. In this paper, we introduce and formalise a process for collecting a FeedbackLog. We also provide concrete use cases where FeedbackLogs can be employed as evidence for algorithmic auditing and as a tool to record updates based on stakeholder feedback.