arXiv subjects

Michael Collins

Publications and source records attributed to Michael Collins.

At least 19 recordsLinked to original sources

arXiv2026

Weird Machines in Transport Layer Security

Weird machines are latent computational capabilities that emerge from the composition of architectural components. Prior work has studied this phenomenon extensively in software systems, including x86 instructions, ELF metadata, and page tables, and more recently in cyber-physical systems such as industrial control networks. This paper extends weird machine theory to a new domain: the Transport Layer Security (TLS) handshake and its two dominant implementations, OpenSSL and BoringSSL. We show that legitimate TLS primitives, including session cache entries, renegotiation logic, extension parsing, and certificate verification steps, compose into Turing-complete systems whose computation is coupled to authentication and trust decisions rather than physical actuation. We formalize this coupling, which we call trust actuation, and argue that any TLS implementation providing session storage, arithmetic on sequence counters, conditional branching on handshake state, and iteration through resumption or retry loops satisfies the conditions for arbitrary computation. We validate this theory with two working demonstrations built on real OpenSSL code paths. The first, a sentinel system, composes standard TLS primitives into a defensive mechanism that detects anomalous handshake behavior. The second, an authentication bypass, composes the same class of primitives into an attack that defeats a cipher-strength policy check through mid-connection renegotiation, without any memory corruption or external malware. Both demonstrations run against real server and client binaries in Docker.

cs.CR↗

arXiv2026

A Sociotechnical, Practitioner-Centered Approach to Technology Adoption in Cybersecurity Operations: An LLM Case

Technology for security operations centers (SOCs) has a storied history of slow adoption due to concerns about trust and reliability. These concerns are amplified with artificial intelligence, particularly large language models (LLMs), which exhibit issues such as hallucinations and inconsistent outputs. To assess whether LLM-based tools can improve SOC efficiency, we embedded two PhD researchers within a multinational company SOC for six months of ethnographic fieldwork. We identified recurring challenges, such as repetitive tasks, fragmented/unclear data, and tooling bottlenecks, and collaborated directly with practitioners to develop LLM companion tools aligned with their operational needs. Iterative refinement reduced workflow disruption and improved interpretability, leading from skepticism to sustained adoption. Ethnographic analysis indicates that this shift was enabled by our sociotechnical co-creation process consistent with Nonaka's SECI model. This framework explains the common challenges in traditional SOC technology adoption, including workflow misalignment, rigidity against evolving threats and internal requirements, and stagnation over time. Our findings show that the co-creation approach can overcome these old barriers and create a new paradigm for creating usable technology for cybersecurity operations.

cs.CR↗

arXiv2025

Hunting in the Dark: Metrics for Early Stage Traffic Discovery

Threat hunting is an operational security process where an expert analyzes traffic, applying knowledge and lightweight tools on unlabeled data in order to identify and classify previously unknown phenomena. In this paper, we examine threat hunting metrics and practice by studying the detection of Crackonosh, a cryptojacking malware package, has on various metrics for identifying its behavior. Using a metric for discoverability, we model the ability of defenders to measure Crackonosh traffic as the malware population decreases, evaluate the strength of various detection methods, and demonstrate how different darkspace sizes affect both the ability to track the malware, but enable emergent behaviors by exploiting attacker mistakes.

cs.CR↗

arXiv2025

Failure of Optimal Design Theory? A Case Study in Toxicology Using Sequential Robust Optimal Design Framework

This paper presents a quasi-sequential optimal design framework for toxicology experiments, specifically applied to sea urchin embryos. The authors propose a novel approach combining robust optimal design with adaptive, stage-based testing to improve efficiency in toxicological studies, particularly where traditional uniform designs fall short. The methodology uses statistical models to refine dose levels across experimental phases, aiming for increased precision while reducing costs and complexity. Key components include selecting an initial design, iterative dose optimization based on preliminary results, and assessing various model fits to ensure robust, data-driven adjustments. Through case studies, we demonstrate improved statistical efficiency and adaptability in toxicology, with potential applications in other experimental domains.

stat.ME↗

arXiv2024

InfAlign: Inference-aware language model alignment

Language model alignment is a critical step in training modern generative language models. Alignment targets to improve win rate of a sample from the aligned model against the base model. Today, we are increasingly using inference-time algorithms (e.g., Best-of-N, controlled decoding, tree search) to decode from language models rather than standard sampling. We show that this train/test mismatch makes standard RLHF framework sub-optimal in view of such inference-time methods. To this end, we propose a framework for inference-aware alignment (InfAlign), which aims to optimize inference-time win rate of the aligned policy against the base model. We prove that for any inference-time decoding procedure, the optimal aligned policy is the solution to the standard RLHF problem with a transformation of the reward. This motivates us to provide the calibrate-and-transform RL (InfAlign-CTRL) algorithm to solve this problem, which involves a reward calibration step and a KL-regularized reward maximization step with a transformation of the calibrated reward. For best-of-N sampling and best-of-N jailbreaking, we propose specific transformations offering up to 3-8% improvement on inference-time win rates. Finally, we also show that our proposed reward calibration method is a strong baseline for optimizing standard win rate.

cs.LG↗

arXiv2024

The Lowell Observatory Solar Telescope: A fiber feed into the EXtreme PREcision Spectrometer

The signal induced by a temperate, terrestrial planet orbiting a Sun-like star is an order of magnitude smaller than the host stars' intrinsic variability. Understanding stellar activity is, therefore, a fundamental obstacle in confirming the smallest exoplanets. We present the Lowell Observatory Solar Telescope (LOST), a solar feed for the EXtreme PREcision Spectrometer (EXPRES) at the 4.3-m Lowell Discovery Telescope (LDT). EXPRES is one of the newest high-resolution spectrographs that accurately measure extreme radial velocity. With LOST/EXPRES, we observe disk-integrated sunlight autonomously throughout the day. In clear conditions, we achieve a ~137,500 optical spectrum of the Sun with a signal-to-noise of 500 in ~150s. Data is reduced using the standard EXPRES pipeline with minimal modification to ensure the data are comparable to the observations of other stars with the LDT. During the first three years of operation, we find a daily RMS of 71 cm/s. Additionally, having two EPRV spectrometers located in Arizona gives us an unprecedented opportunity to benchmark the performance of these planet-finders. We find a RMS of just 55 cm/s when comparing data taken simultaneously with EXPRES and NEID.

astro-ph.EP↗

arXiv2024

Long-Span Question-Answering: Automatic Question Generation and QA-System Ranking via Side-by-Side Evaluation

We explore the use of long-context capabilities in large language models to create synthetic reading comprehension data from entire books. Previous efforts to construct such datasets relied on crowd-sourcing, but the emergence of transformers with a context size of 1 million or more tokens now enables entirely automatic approaches. Our objective is to test the capabilities of LLMs to analyze, understand, and reason over problems that require a detailed comprehension of long spans of text, such as questions involving character arcs, broader themes, or the consequences of early actions later in the story. We propose a holistic pipeline for automatic data generation including question generation, answering, and model scoring using an ``Evaluator''. We find that a relative approach, comparing answers between models in a pairwise fashion and ranking with a Bradley-Terry model, provides a more consistent and differentiating scoring mechanism than an absolute scorer that rates answers individually. We also show that LLMs from different model families produce moderate agreement in their ratings. We ground our approach using the manually curated NarrativeQA dataset, where our evaluator shows excellent agreement with human judgement and even finds errors in the dataset. Using our automatic evaluation approach, we show that using an entire book as context produces superior reading comprehension performance compared to baseline no-context (parametric knowledge only) and retrieval-based approaches.

cs.CL↗

arXiv2024

Superflows: A New Tool for Forensic Network Flow Analysis

Network security analysts gather data from diverse sources, from high-level summaries of network flow and traffic volumes to low-level details such as service logs from servers and the contents of individual packets. They validate and check this data against traffic patterns and historical indicators of compromise. Based on the results of this analysis, a decision is made to either automatically manage the traffic or report it to an analyst for further investigation. Unfortunately, due rapidly increasing traffic volumes, there are far more events to check than operational teams can handle for effective forensic analysis. However, just as packets are grouped into flows that share a commonality, we argue that a high-level construct for grouping network flows into a set a flows that share a hypothesis is needed to significantly improve the quality of operational network response by increasing Events Per Analysts Hour (EPAH). In this paper, we propose a formalism for describing a superflow construct, which we characterize as an aggregation of one or more flows based on an analyst-specific hypothesis about traffic behavior. We demonstrate simple superflow constructions and representations, and perform a case study to explain how the formalism can be used to reduce the volume of data for forensic analysis.

cs.NI↗

arXiv2024

A Chain-of-Thought Is as Strong as Its Weakest Link: A Benchmark for Verifiers of Reasoning Chains

Prompting language models to provide step-by-step answers (e.g., "Chain-of-Thought") is the prominent approach for complex reasoning tasks, where more accurate reasoning chains typically improve downstream task performance. Recent literature discusses automatic methods to verify reasoning to evaluate and improve their correctness. However, no fine-grained step-level datasets are available to enable thorough evaluation of such verification methods, hindering progress in this direction. We introduce REVEAL: Reasoning Verification Evaluation, a dataset to benchmark automatic verifiers of complex Chain-of-Thought reasoning in open-domain question-answering settings. REVEAL includes comprehensive labels for the relevance, attribution to evidence passages, and logical correctness of each reasoning step in a language model's answer, across a variety of datasets and state-of-the-art language models. Evaluation on REVEAL shows that verifiers struggle at verifying reasoning chains - in particular, verifying logical correctness and detecting contradictions. Available at https://reveal-dataset.github.io/ .

cs.CL↗

arXiv2023

Controlled Decoding from Language Models

KL-regularized reinforcement learning (RL) is a popular alignment framework to control the language model responses towards high reward outcomes. We pose a tokenwise RL objective and propose a modular solver for it, called controlled decoding (CD). CD exerts control through a separate prefix scorer module, which is trained to learn a value function for the reward. The prefix scorer is used at inference time to control the generation from a frozen base model, provably sampling from a solution to the RL objective. We empirically demonstrate that CD is effective as a control mechanism on popular benchmarks. We also show that prefix scorers for multiple rewards may be combined at inference time, effectively solving a multi-objective RL problem with no additional training. We show that the benefits of applying CD transfer to an unseen base model with no further tuning as well. Finally, we show that CD can be applied in a blockwise decoding fashion at inference-time, essentially bridging the gap between the popular best-of-K strategy and tokenwise control through reinforcement learning. This makes CD a promising approach for alignment of language models.

cs.LG↗

arXiv2023

Learning to Reject with a Fixed Predictor: Application to Decontextualization

We study the problem of classification with a reject option for a fixed predictor, applicable in natural language processing. We introduce a new problem formulation for this scenario, and an algorithm minimizing a new surrogate loss function. We provide a complete theoretical analysis of the surrogate loss function with a strong $H$-consistency guarantee. For evaluation, we choose the decontextualization task, and provide a manually-labelled dataset of $2\mathord,000$ examples. Our algorithm significantly outperforms the baselines considered, with a $\sim\!\!25\%$ improvement in coverage when halving the error rate, which is only $\sim\!\! 3 \%$ away from the theoretical limit.

cs.LG↗

arXiv2022

Attributed Question Answering: Evaluation and Modeling for Attributed Large Language Models

Large language models (LLMs) have shown impressive results while requiring little or no direct supervision. Further, there is mounting evidence that LLMs may have potential in information-seeking scenarios. We believe the ability of an LLM to attribute the text that it generates is likely to be crucial in this setting. We formulate and study Attributed QA as a key first step in the development of attributed LLMs. We propose a reproducible evaluation framework for the task and benchmark a broad set of architectures. We take human annotations as a gold standard and show that a correlated automatic metric is suitable for development. Our experimental work gives concrete answers to two key questions (How to measure attribution?, and How well do current state-of-the-art methods perform on attribution?), and give some hints as to how to address a third (How to build LLMs with attribution?).

cs.CL↗

arXiv2022

Coreference Resolution through a seq2seq Transition-Based System

Most recent coreference resolution systems use search algorithms over possible spans to identify mentions and resolve coreference. We instead present a coreference resolution system that uses a text-to-text (seq2seq) paradigm to predict mentions and links jointly. We implement the coreference system as a transition system and use multilingual T5 as an underlying language model. We obtain state-of-the-art accuracy on the CoNLL-2012 datasets with 83.3 F1-score for English (a 2.3 higher F1-score than previous work (Dobrovolskii, 2021)) using only CoNLL data for training, 68.5 F1-score for Arabic (+4.1 higher than previous work) and 74.3 F1-score for Chinese (+5.3). In addition we use the SemEval-2010 data sets for experiments in the zero-shot setting, a few-shot setting, and supervised setting using all available training data. We get substantially higher zero-shot F1-scores for 3 out of 4 languages than previous approaches and significantly exceed previous supervised state-of-the-art results for all five tested languages.

cs.CL↗

arXiv2022

Towards Computationally Verifiable Semantic Grounding for Language Models

The paper presents an approach to semantic grounding of language models (LMs) that conceptualizes the LM as a conditional model generating text given a desired semantic message formalized as a set of entity-relationship triples. It embeds the LM in an auto-encoder by feeding its output to a semantic parser whose output is in the same representation domain as the input message. Compared to a baseline that generates text using greedy search, we demonstrate two techniques that improve the fluency and semantic accuracy of the generated text: The first technique samples multiple candidate text sequences from which the semantic parser chooses. The second trains the language model while keeping the semantic parser frozen to improve the semantic accuracy of the auto-encoder. We carry out experiments on the English WebNLG 3.0 data set, using BLEU to measure the fluency of generated text and standard parsing metrics to measure semantic accuracy. We show that our proposed approaches significantly improve on the greedy search baseline. Human evaluation corroborates the results of the automatic evaluation experiments.

cs.CL↗

arXiv2022

Query Refinement Prompts for Closed-Book Long-Form Question Answering

Large language models (LLMs) have been shown to perform well in answering questions and in producing long-form texts, both in few-shot closed-book settings. While the former can be validated using well-known evaluation metrics, the latter is difficult to evaluate. We resolve the difficulties to evaluate long-form output by doing both tasks at once -- to do question answering that requires long-form answers. Such questions tend to be multifaceted, i.e., they may have ambiguities and/or require information from multiple sources. To this end, we define query refinement prompts that encourage LLMs to explicitly express the multifacetedness in questions and generate long-form answers covering multiple facets of the question. Our experiments on two long-form question answering datasets, ASQA and AQuAMuSe, show that using our prompts allows us to outperform fully finetuned models in the closed book setting, as well as achieve results comparable to retrieve-then-generate open-book models.

cs.CL↗

arXiv2022

Improving Low-Resource Cross-lingual Parsing with Expected Statistic Regularization

We present Expected Statistic Regularization (ESR), a novel regularization technique that utilizes low-order multi-task structural statistics to shape model distributions for semi-supervised learning on low-resource datasets. We study ESR in the context of cross-lingual transfer for syntactic analysis (POS tagging and labeled dependency parsing) and present several classes of low-order statistic functions that bear on model behavior. Experimentally, we evaluate the proposed statistics with ESR for unsupervised transfer on 5 diverse target languages and show that all statistics, when estimated accurately, yield improvements to both POS and LAS, with the best statistic improving POS by +7.0 and LAS by +8.5 on average. We also present semi-supervised transfer and learning curve experiments that show ESR provides significant gains over strong cross-lingual-transfer-plus-fine-tuning baselines for modest amounts of label data. These results indicate that ESR is a promising and complementary approach to model-transfer approaches for cross-lingual parsing.

cs.CL↗

arXiv2022

Honest Students from Untrusted Teachers: Learning an Interpretable Question-Answering Pipeline from a Pretrained Language Model

Explainable question answering systems should produce not only accurate answers but also rationales that justify their reasoning and allow humans to check their work. But what sorts of rationales are useful and how can we train systems to produce them? We propose a new style of rationale for open-book question answering, called \emph{markup-and-mask}, which combines aspects of extractive and free-text explanations. In the markup phase, the passage is augmented with free-text markup that enables each sentence to stand on its own outside the discourse context. In the masking phase, a sub-span of the marked-up passage is selected. To train a system to produce markup-and-mask rationales without annotations, we leverage in-context learning. Specifically, we generate silver annotated data by sending a series of prompts to a frozen pretrained language model, which acts as a teacher. We then fine-tune a smaller student model by training on the subset of rationales that led to correct answers. The student is "honest" in the sense that it is a pipeline: the rationale acts as a bottleneck between the passage and the answer, while the "untrusted" teacher operates under no such constraints. Thus, we offer a new way to build trustworthy pipeline systems from a combination of end-task annotations and frozen pretrained language models.

cs.CL↗

arXiv2022

A Well-Composed Text is Half Done! Composition Sampling for Diverse Conditional Generation

We propose Composition Sampling, a simple but effective method to generate diverse outputs for conditional generation of higher quality compared to previous stochastic decoding strategies. It builds on recently proposed plan-based neural generation models (Narayan et al, 2021) that are trained to first create a composition of the output and then generate by conditioning on it and the input. Our approach avoids text degeneration by first sampling a composition in the form of an entity chain and then using beam search to generate the best possible text grounded to this entity chain. Experiments on summarization (CNN/DailyMail and XSum) and question generation (SQuAD), using existing and newly proposed automatic metrics together with human-based evaluation, demonstrate that Composition Sampling is currently the best available decoding strategy for generating diverse meaningful outputs.

cs.CL↗