Search

While methods based on Vision Transformers (ViT) have achieved state-of-the-art performance in many domains, they have not yet been applied successfully in the domain of writer retrieval. The field is dominated by methods using handcrafted features or features extracted from Convolutional Neural Networks. In this work, we bridge this gap and present a novel method that extracts features from a ViT and aggregates them using VLAD encoding. The model is trained in a self-supervised fashion without any need for labels. We show that extracting local foreground features is superior to using the ViT's class token in the context of writer retrieval. We evaluate our method on two historical document collections. We set a new state-at-of-art performance on the Historical-WI dataset (83.1\% mAP), and the HisIR19 dataset (95.0\% mAP). Additionally, we demonstrate that our ViT feature extractor can be directly applied to modern datasets such as the CVL database (98.6\% mAP) without any fine-tuning.

We show that certain moduli spaces of vector bundles over blown-up primary Hopf surfaces admit no compact components. These are the moduli spaces used by Andrei Teleman in his work on the classification of class $VII$ surfaces., Comment: The paper is published

AI systems that serve natural language questions over databases promise to unlock tremendous value. Such systems would allow users to leverage the powerful reasoning and knowledge capabilities of language models (LMs) alongside the scalable computational power of data management systems. These combined capabilities would empower users to ask arbitrary natural language questions over custom data sources. However, existing methods and benchmarks insufficiently explore this setting. Text2SQL methods focus solely on natural language questions that can be expressed in relational algebra, representing a small subset of the questions real users wish to ask. Likewise, Retrieval-Augmented Generation (RAG) considers the limited subset of queries that can be answered with point lookups to one or a few data records within the database. We propose Table-Augmented Generation (TAG), a unified and general-purpose paradigm for answering natural language questions over databases. The TAG model represents a wide range of interactions between the LM and database that have been previously unexplored and creates exciting research opportunities for leveraging the world knowledge and reasoning capabilities of LMs over data. We systematically develop benchmarks to study the TAG problem and find that standard methods answer no more than 20% of queries correctly, confirming the need for further research in this area. We release code for the benchmark at https://github.com/TAG-Research/TAG-Bench.

It is shown that the space of finite-to-finite holomorphic correspondences on an OT-manifold is discrete. When the OT-manifold has no proper infinite complex-analytic subsets, it then follows by known model-theoretic results that its cartesian powers have no interesting complex-analytic families of subvarieties. The methods of proof, which are similar to [Moosa, Moraru, and Toma ``An essentially saturated surface not of K\"ahler-type", {\em Bull. of the LMS}, 40(5):845--854, 2008], require studying finite unramified covers of OT-manifolds., Comment: The paper is published

As practitioners seek to surpass the current reliability and quality frontier of monolithic models, Compound AI Systems consisting of many language model inference calls are increasingly employed. In this work, we construct systems, which we call Networks of Networks (NoNs) organized around the distinction between generating a proposed answer and verifying its correctness, a fundamental concept in complexity theory that we show empirically extends to Language Models (LMs). We introduce a verifier-based judge NoN with K generators, an instantiation of "best-of-K" or "judge-based" compound AI systems. Through experiments on synthetic tasks such as prime factorization, and core benchmarks such as the MMLU, we demonstrate notable performance gains. For instance, in factoring products of two 3-digit primes, a simple NoN improves accuracy from 3.7\% to 36.6\%. On MMLU, a verifier-based judge construction with only 3 generators boosts accuracy over individual GPT-4-Turbo calls by 2.8\%. Our analysis reveals that these gains are most pronounced in domains where verification is notably easier than generation--a characterization which we believe subsumes many reasoning and procedural knowledge tasks, but doesn't often hold for factual and declarative knowledge-based settings. For mathematical and formal logic reasoning-based subjects of MMLU, we observe a 5-8\% or higher gain, whilst no gain on others such as geography and religion. We provide key takeaways for ML practitioners, including the importance of considering verification complexity, the impact of witness format on verifiability, and a simple test to determine the potential benefit of this NoN approach for a given problem distribution. This work aims to inform future research and practice in the design of compound AI systems.

We survey old and new results on the existence of moduli spaces of semistable coherent sheaves both in algebraic and in complex geometry.

The semantic capabilities of language models (LMs) have the potential to enable rich analytics and reasoning over vast knowledge corpora. Unfortunately, existing systems lack high-level abstractions to perform semantic queries at scale. We introduce semantic operators, a declarative programming interface that extends the relational model with composable AI-based operations for semantic queries over datasets (e.g., sorting or aggregating records using natural language criteria). Each operator can be implemented and optimized in multiple ways, opening a rich space for execution plans similar to relational operators. We implement our operators and several optimizations for them in LOTUS, an open-source query engine with a Pandas-like API. We demonstrate LOTUS' effectiveness across a series of real applications, including fact-checking, extreme multi-label classification, and search. We find that LOTUS' programming model is highly expressive, capturing state-of-the-art query pipelines with low development overhead. Specifically, on the FEVER dataset, LOTUS' programs can reproduce FacTool, a recent state-of-the-art fact-checking pipeline, in few lines of code, and implement a new pipeline that improves accuracy by $9.5\%$, while offering $7-34\times$ lower execution time. In the extreme multi-label classification task on the BioDEX dataset, LOTUS reproduces state-of-the art result quality with its join operator, while providing an efficient algorithm that runs $800\times$ faster than a naive join. In the search and ranking application, LOTUS allows a simple composition of operators to achieve $5.9 - 49.4\%$ higher nDCG@10 than the vanilla retriever and re-ranker, while also providing query efficiency, with $1.67 - 10\times$ lower execution time than LM-based ranking methods used by prior works. LOTUS is publicly available at https://github.com/stanford-futuredata/lotus.

We investigate the influence of boundary terms in the warm inflationary scenario, by considering that in the Einstein-Hilbert action the boundary can be described in terms of a Weyl-type geometry. The gravitational action, as well as the field equations, are thus extended to include new geometrical terms, coming from the non-metric nature of the boundary, and depending on the Weyl vector, and its covariant derivatives. We investigate the effects of these new boundary terms by considering the warm inflationary scenario of the early evolution of the Universe, in the presence of a scalar field. We obtain the generalized Friedmann equations in the Universe with a Weylian boundary by considering the Friedmann-Lemaitre-Robertson-Walker metric. We consider the simultaneous decay of the scalar field, and of the creation of radiation, by appropriately splitting the general conservation equation through the introduction of the dissipation coefficient, which can depend on both the scalar field, and the Weyl vector. We consider three distinct warm inflationary models, in which the dissipation coefficients are chosen as different functions of the scalar field and of the Weyl vector. The numerical solutions of the cosmological evolution equations show that the radiation is created during the very early phases of expansion, and, after the radiation reaches its maximum value, the transition from an accelerating inflationary phase to a decelerating one takes place. Moreover, it turns out that the the Weyl vector, describing the boundary effects on the cosmological evolution, plays a significant role during the process of radiation creation., Comment: 18 pages, 6 figures, accepted for publication in PDU

Large Language Models (LLMs) have had considerable difficulty when prompted with mathematical questions, especially those within theory of computing (ToC) courses. In this paper, we detail two experiments regarding our own ToC course and the ChatGPT LLM. For the first, we evaluated ChatGPT's ability to pass our own ToC course's exams. For the second, we created a database of sample ToC questions and responses to accommodate other ToC offerings' choices for topics and structure. We scored each of ChatGPT's outputs on these questions. Overall, we determined that ChatGPT can pass our ToC course, and is adequate at understanding common formal definitions and answering "simple"-style questions, e.g., true/false and multiple choice. However, ChatGPT often makes nonsensical claims in open-ended responses, such as proofs., Comment: Accepted to SIGCSE Virtual 2024

We study the moduli stacks of slope-semistable torsion-free coherent sheaves that admit reflexive, respectively locally free, Seshadri graduations on a smooth projective variety. We show that they are open in the stack of coherent sheaves and that they admit good moduli spaces when the field characteristic is zero. In addition, in the locally free case we prove that the resulting moduli space is a quasi-projective scheme., Comment: v2: Added remarks 3.1 and 3.2 on openness in the analytic case, small notational change

Intent inferral on a hand orthosis for stroke patients is challenging due to the difficulty of data collection from impaired subjects. Additionally, EMG signals exhibit significant variations across different conditions, sessions, and subjects, making it hard for classifiers to generalize. Traditional approaches require a large labeled dataset from the new condition, session, or subject to train intent classifiers; however, this data collection process is burdensome and time-consuming. In this paper, we propose ChatEMG, an autoregressive generative model that can generate synthetic EMG signals conditioned on prompts (i.e., a given sequence of EMG signals). ChatEMG enables us to collect only a small dataset from the new condition, session, or subject and expand it with synthetic samples conditioned on prompts from this new context. ChatEMG leverages a vast repository of previous data via generative training while still remaining context-specific via prompting. Our experiments show that these synthetic samples are classifier-agnostic and can improve intent inferral accuracy for different types of classifiers. We demonstrate that our complete approach can be integrated into a single patient session, including the use of the classifier for functional orthosis-assisted tasks. To the best of our knowledge, this is the first time an intent classifier trained partially on synthetic data has been deployed for functional control of an orthosis by a stroke survivor. Videos and additional information can be found at https://jxu.ai/chatemg., Comment: 8 pages

Language Model Programs, i.e. sophisticated pipelines of modular language model (LM) calls, are increasingly advancing NLP tasks, but they require crafting prompts that are jointly effective for all modules. We study prompt optimization for LM programs, i.e. how to update these prompts to maximize a downstream metric without access to module-level labels or gradients. To make this tractable, we factorize our problem into optimizing the free-form instructions and few-shot demonstrations of every module and introduce several strategies to craft task-grounded instructions and navigate credit assignment across modules. Our strategies include (i) program- and data-aware techniques for proposing effective instructions, (ii) a stochastic mini-batch evaluation function for learning a surrogate model of our objective, and (iii) a meta-optimization procedure in which we refine how LMs construct proposals over time. Using these insights we develop MIPRO, a novel optimizer that outperforms baselines on five of six diverse LM programs using a best-in-class open-source model (Llama-3-8B), by as high as 12.9% accuracy. We will release our new optimizers and benchmark in DSPy at https://github.com/stanfordnlp/dspy, Comment: Krista and Michael contributed equally to this work

Underactuated manipulators reduce the number of bulky motors, thereby enabling compact and mechanically robust designs. However, fewer actuators than joints means that the manipulator can only access a specific manifold within the joint space, which is particular to a given hardware configuration and can be low-dimensional and/or discontinuous. Determining an appropriate set of hardware parameters for this class of mechanisms, therefore, is difficult - even for traditional task-based co-optimization methods. In this paper, our goal is to implement a task-based design and policy co-optimization method for underactuated, tendon-driven manipulators. We first formulate a general model for an underactuated, tendon-driven transmission. We then use this model to co-optimize a three-link, two-actuator kinematic chain using reinforcement learning. We demonstrate that our optimized tendon transmission and control policy can be transferred reliably to physical hardware with real-world reaching experiments.

Let $f(t_1, \ldots, t_r, X)\in \mathbb{Z}[t_1, \ldots, t_r,X]$ be irreducible and let $a_1, \ldots, a_r\in \mathbb{Z} \smallsetminus \{0,\pm 1\}$. Under a necessary ramification assumption on $f$, and conditionally on the Generalized Riemann Hypothesis, we show that for almost all integers $n_1, \ldots, n_r$, the polynomial $f(a_1^{n_1}, \ldots, a_r^{n_r}, X)$ is irreducible in $\mathbb{Q}[X]$.

For cyber-physical systems (CPS), including robotics and autonomous vehicles, mass deployment has been hindered by fatal errors that occur when operating in rare events. To replicate rare events such as vehicle crashes, many companies have created logging systems and employed crash reconstruction experts to meticulously recreate these valuable events in simulation. However, in these methods, "what if" questions are not easily formulated and answered. We present ScenarioNL, an AI System for creating scenario programs from natural language. Specifically, we generate these programs from police crash reports. Reports normally contain uncertainty about the exact details of the incidents which we represent through a Probabilistic Programming Language (PPL), Scenic. By using Scenic, we can clearly and concisely represent uncertainty and variation over CPS behaviors, properties, and interactions. We demonstrate how commonplace prompting techniques with the best Large Language Models (LLM) are incapable of reasoning about probabilistic scenario programs and generating code for low-resource languages such as Scenic. Our system is comprised of several LLMs chained together with several kinds of prompting strategies, a compiler, and a simulator. We evaluate our system on publicly available autonomous vehicle crash reports in California from the last five years and share insights into how we generate code that is both semantically meaningful and syntactically correct., Comment: 17 pages, 2 figures

This work introduces HistNERo, the first Romanian corpus for Named Entity Recognition (NER) in historical newspapers. The dataset contains 323k tokens of text, covering more than half of the 19th century (i.e., 1817) until the late part of the 20th century (i.e., 1990). Eight native Romanian speakers annotated the dataset with five named entities. The samples belong to one of the following four historical regions of Romania, namely Bessarabia, Moldavia, Transylvania, and Wallachia. We employed this proposed dataset to perform several experiments for NER using Romanian pre-trained language models. Our results show that the best model achieved a strict F1-score of 55.69%. Also, by reducing the discrepancies between regions through a novel domain adaption technique, we improved the performance on this corpus to a strict F1-score of 66.80%, representing an absolute gain of more than 10%., Comment: Accepted at the International Conference on Document Analysis and Recognition (ICDAR 2024)

Robotic systems are becoming pervasive and adopted in increasingly many domains, such as manufacturing, healthcare, and space exploration. To this end, engineering software has emerged as a crucial discipline for building maintainable and reusable robotic systems. The robotics software engineering research field has received increasing attention, fostering autonomy as a fundamental goal. However, robotics developers are still challenged to achieve this goal because simulation cannot realistically deliver solutions to emulate real-world phenomena. Robots also need to operate in unpredictable and uncontrollable environments, which require safe and trustworthy self-adaptation capabilities implemented in software. Typical techniques to address the challenges are runtime verification, field-based testing, and mitigation techniques that enable fail-safe solutions. However, no clear guidance exists for architecting ROS-based systems to enable and facilitate runtime verification and field-based testing. This paper aims to fill this gap by providing guidelines to help developers and quality assurance (QA) teams develop, verify, or test their robots in the field. These guidelines are carefully tailored to address the challenges and requirements of testing robotics systems in real-world scenarios. We conducted (i) a literature review on studies addressing runtime verification and field-based testing for robotic systems, (ii) mined ROS-based applications repositories, and (iii) validated the applicability, clarity, and usefulness via two questionnaires with 55 answers overall. We contribute 20 guidelines: 8 for developers and 12 for QA teams formulated for researchers and practitioners in robotic software engineering. Finally, we map our guidelines to open challenges in runtime verification and field-based testing for ROS-based systems, and we outline promising research directions in the field.

We propose MetaEMG, a meta-learning approach for fast adaptation in intent inferral on a robotic hand orthosis for stroke. One key challenge in machine learning for assistive and rehabilitative robotics with disabled-bodied subjects is the difficulty of collecting labeled training data. Muscle tone and spasticity often vary significantly among stroke subjects, and hand function can even change across different use sessions of the device for the same subject. We investigate the use of meta-learning to mitigate the burden of data collection needed to adapt high-capacity neural networks to a new session or subject. Our experiments on real clinical data collected from five stroke subjects show that MetaEMG can improve the intent inferral accuracy with a small session- or subject-specific dataset and very few fine-tuning epochs. To the best of our knowledge, we are the first to formulate intent inferral on stroke subjects as a meta-learning problem and demonstrate fast adaptation to a new session or subject for controlling a robotic hand orthosis with EMG signals.

We prove uniform boundedness statements for semistable pure sheaves on projective manifolds. For example, we prove that the set of isomorphism classes of pure sheaves of dimension 2 that are slope semistable with respect to ample classes that vary in a compact set $K$ are bounded. We also prove uniform boundedness for pure sheaves of higher dimension, but with restrictions on the compact set $K$. As applications we get new statements about moduli spaces of semistable sheaves, and the wall-chamber structure that governs their variation., Comment: 24 pages, 1 figure

Pretraining Large Language Models (LLMs) on large corpora of textual data is now a standard paradigm. When using these LLMs for many downstream applications, it is common to additionally bake in new knowledge (e.g., time-critical news, or private domain knowledge) into the pretrained model either through RAG-based-prompting, or fine-tuning. However, the optimal methodology for the model to gain such new knowledge remains an open question. In this paper, we present Retrieval Augmented FineTuning (RAFT), a training recipe that improves the model's ability to answer questions in a "open-book" in-domain settings. In RAFT, given a question, and a set of retrieved documents, we train the model to ignore those documents that don't help in answering the question, which we call, distractor documents. RAFT accomplishes this by citing verbatim the right sequence from the relevant document that would help answer the question. This coupled with RAFT's chain-of-thought-style response helps improve the model's ability to reason. In domain-specific RAG, RAFT consistently improves the model's performance across PubMed, HotpotQA, and Gorilla datasets, presenting a post-training recipe to improve pre-trained LLMs to in-domain RAG. RAFT's code and demo are open-sourced at github.com/ShishirPatil/gorilla.

Analytical database providers (e.g., Redshift, Databricks, BigQuery) have rapidly added support for invoking Large Language Models (LLMs) through native user-defined functions (UDFs) to help users perform natural language tasks, such as classification, entity extraction, and translation, inside analytical workloads. For instance, an analyst might want to extract customer sentiments on millions of product reviews. However, LLM inference is highly expensive in both computational and economic terms: for example, an NVIDIA L4 GPU running Llama2-7B can only process 6 KB of text per second. In this paper, we explore how to optimize LLM inference for analytical workloads that invoke LLMs within relational queries. We show that relational queries present novel opportunities for accelerating LLM inference, including reordering rows to maximize key-value (KV) cache reuse within the LLM inference engine, reordering columns within a row to further increase cache reuse, and deduplicating redundant inference requests. We implement these optimizations in Apache Spark, with vLLM as the model serving backend and achieve up to 4.4x improvement in end-to-end latency on a benchmark of diverse LLM-based queries on real datasets. To the best of our knowledge, this is the first work to explicitly address the problem of optimizing LLM invocations within SQL queries.

We present ALTO, a network orchestrator for efficiently serving compound AI systems such as pipelines of language models. ALTO achieves high throughput and low latency by taking advantage of an optimization opportunity specific to generative language models: streaming intermediate outputs. As language models produce outputs token by token, ALTO exposes opportunities to stream intermediate outputs between stages when possible. We highlight two new challenges of correctness and load balancing which emerge when streaming intermediate data across distributed pipeline stage instances. We also motivate the need for an aggregation-aware routing interface and distributed prompt-aware scheduling to address these challenges. We demonstrate the impact of ALTO's partial output streaming on a complex chatbot verification pipeline, increasing throughput by up to 3x for a fixed latency target of 4 seconds / request while also reducing tail latency by 1.8x compared to a baseline serving approach.

Applications increasingly leverage mixed-modality data, and must jointly search over vector data, such as embedded images, text and video, as well as structured data, such as attributes and keywords. Proposed methods for this hybrid search setting either suffer from poor performance or support a severely restricted set of search predicates (e.g., only small sets of equality predicates), making them impractical for many applications. To address this, we present ACORN, an approach for performant and predicate-agnostic hybrid search. ACORN builds on Hierarchical Navigable Small Worlds (HNSW), a state-of-the-art graph-based approximate nearest neighbor index, and can be implemented efficiently by extending existing HNSW libraries. ACORN introduces the idea of predicate subgraph traversal to emulate a theoretically ideal, but impractical, hybrid search strategy. ACORN's predicate-agnostic construction algorithm is designed to enable this effective search strategy, while supporting a wide array of predicate sets and query semantics. We systematically evaluate ACORN on both prior benchmark datasets, with simple, low-cardinality predicate sets, and complex multi-modal datasets not supported by prior methods. We show that ACORN achieves state-of-the-art performance on all datasets, outperforming prior methods with 2-1,000x higher throughput at a fixed recall.

Many recent state-of-the-art results in language tasks were achieved using compound systems that perform multiple Language Model (LM) calls and aggregate their responses. However, there is little understanding of how the number of LM calls - e.g., when asking the LM to answer each question multiple times and taking a majority vote - affects such a compound system's performance. In this paper, we initiate the study of scaling properties of compound inference systems. We analyze, theoretically and empirically, how the number of LM calls affects the performance of Vote and Filter-Vote, two of the simplest compound system designs, which aggregate LM responses via majority voting, optionally applying LM filters. We find, surprisingly, that across multiple language tasks, the performance of both Vote and Filter-Vote can first increase but then decrease as a function of the number of LM calls. Our theoretical results suggest that this non-monotonicity is due to the diversity of query difficulties within a task: more LM calls lead to higher performance on "easy" queries, but lower performance on "hard" queries, and non-monotone behavior can emerge when a task contains both types of queries. This insight then allows us to compute, from a small number of samples, the number of LM calls that maximizes system performance, and define an analytical scaling model for both systems. Experiments show that our scaling model can accurately predict the performance of Vote and Filter-Vote systems and thus find the optimal number of LM calls to make.

Current language models fall short in understanding aspects of the world not easily described in words, and struggle with complex, long-form tasks. Video sequences offer valuable temporal information absent in language and static images, making them attractive for joint modeling with language. Such models could develop a understanding of both human textual knowledge and the physical world, enabling broader AI capabilities for assisting humans. However, learning from millions of tokens of video and language sequences poses challenges due to memory constraints, computational complexity, and limited datasets. To address these challenges, we curate a large dataset of diverse videos and books, utilize the Blockwise RingAttention technique to scalably train on long sequences, and gradually increase context size from 4K to 1M tokens. This paper makes the following contributions: (a) Largest context size neural network: We train one of the largest context size transformers on long video and language sequences, setting new benchmarks in difficult retrieval tasks and long video understanding. (b) Solutions for overcoming vision-language training challenges, including using masked sequence packing for mixing different sequence lengths, loss weighting to balance language and vision, and model-generated QA dataset for long sequence chat. (c) A highly-optimized implementation with RingAttention, Blockwise Transformers, masked sequence packing, and other key features for training on millions-length multimodal sequences. (d) Fully open-sourced a family of 7B parameter models capable of processing long text documents (LWM-Text, LWM-Text-Chat) and videos (LWM, LWM-Chat) of over 1M tokens. This work paves the way for training on massive datasets of long video and language to develop understanding of both human knowledge and the multimodal world, and broader capabilities.

We study a class of semistability conditions defined by a system of ample classes for coherent sheaves over a smooth projective variety. Under some necessary boundedness assumptions, we show the existence of a well-behaved chamber structure for the variation of moduli spaces of sheaves with respect to the change of semistability., Comment: 21 pages, v2: Mistake corrected in the proof of the HN property

Individuals with hand paralysis resulting from C6-C7 spinal cord injuries frequently rely on tenodesis for grasping. However, tenodesis generates limited grasping force and demands constant exertion to maintain a grasp, leading to fatigue and sometimes pain. We introduce the MyHand-SCI, a wearable robot that provides grasping assistance through motorized exotendons. Our user-driven device enables independent, ipsilateral operation via a novel Throttle-based Wrist Angle control method, which allows users to maintain grasps without continued wrist extension. A pilot case study with a person with C6 spinal cord injury shows an improvement in functional grasping and grasping force, as well as a preserved ability to modulate grasping force while using our device, thus improving their ability to manipulate everyday objects. This research is a step towards developing effective and intuitive wearable assistive devices for individuals with spinal cord injury., Comment: 6 pages, 8 figures, 1 table

Increased effort during use of the paretic arm and hand can provoke involuntary abnormal synergy patterns and amplify stiffness effects of muscle tone for individuals after stroke, which can add difficulty for user-controlled devices to assist hand movement during functional tasks. We study how volitional effort, exerted in an attempt to open or close the hand, affects resistance to robot-assisted movement at the finger level. We perform experiments with three chronic stroke survivors to measure changes in stiffness when the user is actively exerting effort to activate ipsilateral EMG-controlled robot-assisted hand movements, compared with when the fingers are passively stretched, as well as overall effects from sustained active engagement and use. Our results suggest that active engagement of the upper extremity increases muscle tone in the finger to a much greater degree than through passive-stretch or sustained exertion over time. Potential design implications of this work suggest that developers should anticipate higher levels of finger stiffness when relying on user-driven ipsilateral control methods for assistive or rehabilitative devices for stroke., Comment: 6 pages, 6 figures, 2 tables. Accepted to BioRob 2024

We introduce GEOTACT, a robotic manipulation method capable of retrieving objects buried in granular media. This is a challenging task due to the need to interact with granular media, and doing so based exclusively on tactile feedback, since a buried object can be completely hidden from vision. Tactile feedback is in itself challenging in this context, due to ubiquitous contact with the surrounding media, and the inherent noise level induced by the tactile readings. To address these challenges, we use a learning method trained end-to-end with simulated sensor noise. We show that our problem formulation leads to the natural emergence of learned pushing behaviors that the manipulator uses to reduce uncertainty and funnel the object to a stable grasp despite spurious and noisy tactile readings. We also introduce a training curriculum that enables learning these behaviors in simulation, followed by zero-shot transfer to real hardware. To the best of our knowledge, GEOTACT is the first method to reliably retrieve a number of different objects from a granular environment, doing so on real hardware and with integrated tactile sensing. Videos and additional information can be found at https://jxu.ai/geotact.

With Artificial Intelligence (AI) influencing the decision-making process of sensitive applications such as Face Verification, it is fundamental to ensure the transparency, fairness, and accountability of decisions. Although Explainable Artificial Intelligence (XAI) techniques exist to clarify AI decisions, it is equally important to provide interpretability of these decisions to humans. In this paper, we present an approach to combine computer and human vision to increase the explanation's interpretability of a face verification algorithm. In particular, we are inspired by the human perceptual process to understand how machines perceive face's human-semantic areas during face comparison tasks. We use Mediapipe, which provides a segmentation technique that identifies distinct human-semantic facial regions, enabling the machine's perception analysis. Additionally, we adapted two model-agnostic algorithms to provide human-interpretable insights into the decision-making processes.

We present a method for enabling Reinforcement Learning of motor control policies for complex skills such as dexterous manipulation. We posit that a key difficulty for training such policies is the difficulty of exploring the problem state space, as the accessible and useful regions of this space form a complex structure along manifolds of the original high-dimensional state space. This work presents a method to enable and support exploration with Sampling-based Planning. We use a generally applicable non-holonomic Rapidly-exploring Random Trees algorithm and present multiple methods to use the resulting structure to bootstrap model-free Reinforcement Learning. Our method is effective at learning various challenging dexterous motor control skills of higher difficulty than previously shown. In particular, we achieve dexterous in-hand manipulation of complex objects while simultaneously securing the object without the use of passive support surfaces. These policies also transfer effectively to real robots. A number of example videos can also be found on the project website: https://sbrl.cs.columbia.edu, Comment: 20 pages, 14 figures, submitted to Autonomous Robots, RSS 2023 Special Issue. arXiv admin note: substantial text overlap with arXiv:2303.03486

In November 2022, OpenAI has introduced ChatGPT, a chatbot based on supervised and reinforcement learning. Not only can it answer questions emulating human-like responses, but it can also generate code from scratch or complete coding templates provided by the user. ChatGPT can generate unique responses which render any traditional anti-plagiarism tool useless. Its release has ignited a heated debate about its usage in academia, especially by students. We have found, to our surprise, that our students at POLITEHNICA University of Bucharest (UPB) have been using generative AI tools (ChatGPT and its predecessors) for solving homework, for at least 6 months. We therefore set out to explore the capabilities of ChatGPT and assess its value for educational purposes. We solved all our coding assignments for the semester from our UPB Functional Programming course. We discovered that, although ChatGPT provides correct answers in 68% of the cases, only around half of those are legible solutions which can benefit students in some form. On the other hand, ChatGPT has a very good ability to perform code review on student programming homework. Based on these findings, we discuss the pros and cons of ChatGPT in education.

The ensemble Kalman inversion (EKI), a recently introduced optimisation method for solving inverse problems, is widely employed for the efficient and derivative-free estimation of unknown parameters. Specifically in cases involving ill-posed inverse problems and high-dimensional parameter spaces, the scheme has shown promising success. However, in its general form, the EKI does not take constraints into account, which are essential and often stem from physical limitations or specific requirements. Based on a log-barrier approach, we suggest adapting the continuous-time formulation of EKI to incorporate convex inequality constraints. We underpin this adaptation with a theoretical analysis that provides lower and upper bounds on the ensemble collapse, as well as convergence to the constraint optimum for general nonlinear forward models. Finally, we showcase our results through two examples involving partial differential equations (PDEs).

Chaining language model (LM) calls as composable modules is fueling a new way of programming, but ensuring LMs adhere to important constraints requires heuristic "prompt engineering". We introduce LM Assertions, a programming construct for expressing computational constraints that LMs should satisfy. We integrate our constructs into the recent DSPy programming model for LMs, and present new strategies that allow DSPy to compile programs with LM Assertions into more reliable and accurate systems. We also propose strategies to use assertions at inference time for automatic self-refinement with LMs. We report on four diverse case studies for text generation and find that LM Assertions improve not only compliance with imposed rules but also downstream task performance, passing constraints up to 164% more often and generating up to 37% more higher-quality responses. Our reference implementation of LM Assertions is integrated into DSPy at https://github.com/stanfordnlp/dspy, Comment: Arnav*, Manish*, Shangyin* contributed equally to this work

Spiking neural networks (SNNs) take inspiration from the brain to enable energy-efficient computations. Since the advent of Transformers, SNNs have struggled to compete with artificial networks on modern sequential tasks, as they inherit limitations from recurrent neural networks (RNNs), with the added challenge of training with non-differentiable binary spiking activations. However, a recent renewed interest in efficient alternatives to Transformers has given rise to state-of-the-art recurrent architectures named state space models (SSMs). This work systematically investigates, for the first time, the intersection of state-of-the-art SSMs with SNNs for long-range sequence modelling. Results suggest that SSM-based SNNs can outperform the Transformer on all tasks of a well-established long-range sequence modelling benchmark. It is also shown that SSM-based SNNs can outperform current state-of-the-art SNNs with fewer parameters on sequential image classification. Finally, a novel feature mixing layer is introduced, improving SNN accuracy while challenging assumptions about the role of binary activations in SNNs. This work paves the way for deploying powerful SSM-based architectures, such as large language models, to neuromorphic hardware for energy-efficient long-range sequence modelling., Comment: 18 pages, 10 Figures/Tables

This paper addresses the challenging task of guide wire navigation in cardiovascular interventions, focusing on the parameter estimation of a guide wire system using Ensemble Kalman Inversion (EKI) with a subsampling technique. The EKI uses an ensemble of particles to estimate the unknown quantities. However since the data misfit has to be computed for each particle in each iteration, the EKI may become computationally infeasible in the case of high-dimensional data, e.g. high-resolution images. This issue can been addressed by randomised algorithms that utilize only a random subset of the data in each iteration. We introduce and analyse a subsampling technique for the EKI, which is based on a continuous-time representation of stochastic gradient methods and apply it to on the parameter estimation of our guide wire system. Numerical experiments with real data from a simplified test setting demonstrate the potential of the method.

The integration of artificial intelligence into scientific research has reached a new pinnacle with GPT-4V, a large language model featuring enhanced vision capabilities, accessible through ChatGPT or an API. This study demonstrates the remarkable ability of GPT-4V to navigate and obtain complex data for metal-organic frameworks, especially from graphical sources. Our approach involved an automated process of converting 346 scholarly articles into 6240 images, which represents a benchmark dataset in this task, followed by deploying GPT-4V to categorize and analyze these images using natural language prompts. This methodology enabled GPT-4V to accurately identify and interpret key plots integral to MOF characterization, such as nitrogen isotherms, PXRD patterns, and TGA curves, among others, with accuracy and recall above 93%. The model's proficiency in extracting critical information from these plots not only underscores its capability in data mining but also highlights its potential in aiding the creation of comprehensive digital databases for reticular chemistry. In addition, the extracted nitrogen isotherm data from the selected literature allowed for a comparison between theoretical and experimental porosity values for over 200 compounds, highlighting certain discrepancies and underscoring the importance of integrating computational and experimental data. This work highlights the potential of AI in accelerating scientific discovery and innovation, bridging the gap between computational tools and experimental research, and paving the way for more efficient, inclusive, and comprehensive scientific inquiry., Comment: 36 pages, 24 figures

We address the problem of shock induced ignition and transition to detonation in a reactive medium in the presence of mechanically induced fluctuations by a moving oscillating piston. For the inert problem prior to ignition, we provide a closed form model for the generation of the train of compression and expansions, their steepening into a train of N-shock waves and their reflection on the lead shock, as well as the distribution energy dissipation rate in the induction zone. The model is found in excellent agreement with numerics. Reactive calculations were performed for hydrogen and ethylene fuels using a novel high-fidelity numerical Lagrangian scheme. Different regimes of ignition and transition to detonation, controlled by the time scale of the forcing and the two time scales of the chemistry: the induction and reaction times. Two novel hot spot cascade mechanisms were identified. The first relies on the coherence between the sequence of hot spot formation set by the piston forcing and forward wave interaction with the lead shock, generalizing the classic runaway in fast flames. The second hot spot cascade is triggered by the feedback between the pressure pulse generated by the first generation hot spot cascade and the shock. For slow forcing, the sensitization is through a modification to the classic run-away process, while the high frequency regime leads to very localized sub-critical hot-spot formation controlled by the cumulative energy dissipation of the first generation shocks at a distance comparable to the shock formation location.

The objective of this study is to determine the critical conditions for a detonation wave formation following a Mach reflection of two incident shocks. This problem is central to the propagation mechanism of cellular detonations, where such periodic reflections occur continuously. A new experimental technique is introduced that permits the isolation of this Mach reflection. The technique is inspired by that of White (1963) and uses a detonation passing through a bifurcated converging-diverging nozzle to reproducibly generate the desired shock reflection at its exit, where the two transmitted mis-aligned incident shocks interact. The collision process was monitored with high-speed Schlieren videos. The experiments were performed in mixtures of methane-oxygen and hydrogen-oxygen-argon. Three distinct regimes of transmission were identified: detonation formation, decaying Mach shock followed by a flame, and inert Mach shock with no re-ignition. The condition separating the re-ignition and no re-ignition was very well predicted by the critical decay rate ignition theory of Eckett et al. On the other hand, the transition between detonation formation and ignition was found to correlate much better with the critical curvature concept of quasi-steady curved detonations of Kasimov and Stewart. Both mixtures were found in excellent agreement with this criterion, established on the basis of experimental velocity-curvature data available in the literature for these mixtures. For the hydrogen detonations, this criterion was also in agreement with the critical curvature predicted by ZND theory. For the methane detonations, order of magnitude disagreement was observed with the ZND model prediction, further confirming the propensity of irregular cellular detonations to propagate under higher curvatures than predicted by laminar detonation theory.

A minimal model for curved detonations is studied, illustrating the role of the reaction rate on the detonation speed and its propagation limits. The model is based on a simple extension of the minimal Fickett toy model for detonations based on the kinematic wave equation. The use of a simple depletion rate conditioned on the shock speed serves to illustrate its role in the quasi-steady structure of curved waves and their initiation from a strong blast wave. Calculations of strong initiation from a self-similar explosion illustrate the various asymptotic regimes of the transition to self-sustenance and their link to the steady wave structure. We recover the asymptotic regimes of detonation formation suggested by He and Clavin and modelled in the context of Detonation Shock Dynamics by Stewart and collaborators. Following an analysis using the shock change equation, we identify a unique criterion that permits to infer the critical energy for initiation from the competition between energy release and geometric decay.

As Machine Learning (ML) systems continue to grow, the demand for relevant and comprehensive datasets becomes imperative. There is limited study on the challenges of data acquisition due to ad-hoc processes and lack of consistent methodologies. We first present an investigation of current data marketplaces, revealing lack of platforms offering detailed information about datasets, transparent pricing, standardized data formats. With the objective of inciting participation from the data-centric AI community, we then introduce the DAM challenge, a benchmark to model the interaction between the data providers and acquirers. The benchmark was released as a part of DataPerf. Our evaluation of the submitted strategies underlines the need for effective data acquisition strategies in ML.

We derive the shock strength area rule for a Noble-Abel stiffened gas (NASG) equation of a state required in Whitham's geometrical shock dynamics approach to determine shock wave dynamics in dense gases, liquids and solids. An exact formulation requiring the solution of an ordinary differential equation is provided. Closed form solutions of various levels of approximations are also obtained as an expansion in shock wave strength. The leading order approximation recovers the geometrical acoustic limit, while higher order approximations account for the medium's compressibility. The exact shock strength area relation and the various order approximations are illustrated for shocks in liquid water. The simple closed form of the first order solution predicts the shock strength area rule up to shock pressures of approximately twice the stiffening pressure in water, i.e., approximately 1-2 GPa., Comment: Paper presented at the 34th International Symposium on Shock Waves, 16-21 July 2023, Daegu, South Korea; paper T16-0375