PyTorchKR Papers

최신 논문

총 1,461편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:2306.07703공식 코드 · Papers with Code

E2E-LOAD: End-to-End Long-form Online Action Detection

Recently, there has been a growing trend toward feature-based approaches for Online Action Detection (OAD). However, these approaches have limitations due to their fixed backbone design, which ignores the potential capability of a trainable backbone. In this paper, we propose the first end-to-end OAD model, termed E2E-LOAD, designed to address the major challenge of OAD, namely, long-term understanding and efficient online reasoning. Specifically, our proposed approach adopts an initial spatial model that is shared by all frames and maintains a long sequence cache for inference at a low computational cost. We also advocate an asymmetric spatial-temporal model for long-form and short-form modeling effectively. Furthermore, we propose a novel and efficient inference mechanism that accelerates heavy spatial-temporal exploration. Extensive ablation studies and experiments demonstrate the effectiveness and efficiency of our proposed method. Notably, we achieve 17.3 (+12.6) FPS for end-to-end OAD with 72.4%~(+1.2%), 90.3%~(+0.7%), and 48.1%~(+26.0%) mAP on THMOUS14, TVSeries, and HDD, respectively, which is 3x faster than previous approaches. The source code will be made publicly available.

arXiv:2306.05014공식 코드 · Papers with Code

Learning Closed-form Equations for Subgrid-scale Closures from High-fidelity Data: Promises and Challenges

There is growing interest in discovering interpretable, closed-form equations for subgrid-scale (SGS) closures/parameterizations of complex processes in Earth systems. Here, we apply a common equation-discovery technique with expansive libraries to learn closures from filtered direct numerical simulations of 2D turbulence and Rayleigh-B\'enard convection (RBC). Across common filters (e.g., Gaussian, box), we robustly discover closures of the same form for momentum and heat fluxes. These closures depend on nonlinear combinations of gradients of filtered variables, with constants that are independent of the fluid/flow properties and only depend on filter type/size. We show that these closures are the nonlinear gradient model (NGM), which is derivable analytically using Taylor-series. Indeed, we suggest that with common (physics-free) equation-discovery algorithms, for many common systems/physics, discovered closures are consistent with the leading term of the Taylor-series (except when cutoff filters are used). Like previous studies, we find that large-eddy simulations with NGM closures are unstable, despite significant similarities between the true and NGM-predicted fluxes (correlations $> 0.95$). We identify two shortcomings as reasons for these instabilities: in 2D, NGM produces zero kinetic energy transfer between resolved and subgrid scales, lacking both diffusion and backscattering. In RBC, potential energy backscattering is poorly predicted. Moreover, we show that SGS fluxes diagnosed from data, presumed the ''truth'' for discovery, depend on filtering procedures and are not unique. Accordingly, to learn accurate, stable closures in future work, we propose several ideas around using physics-informed libraries, loss functions, and metrics. These findings are relevant to closure modeling of any multi-scale system.

arXiv:2306.04065

Sustainability criterion implied externality pricing for resource extraction

A dynamic model is constructed that generalises the Hartwick and Van Long (2020) endogenous discounting setup by introducing externalities and asks what implications this has for optimal natural resource extraction with constant consumption. It is shown that a modified form of the Hotelling and Hartwick rule holds in which the externality component of price is a specific function of the instantaneous user costs and cross price elasticities. It is demonstrated that the externality adjusted marginal user cost of remaining natural reserves is equal to the marginal user cost of extracted resources invested in human-made reproducible capital. This lends itself to a discrete form with a readily intuitive economic interpretation that illuminates the stepwise impact of externality pricing on optimal extraction schedules.

arXiv:2306.04537

Long-form analogies generated by chatGPT lack human-like psycholinguistic properties

Psycholinguistic analyses provide a means of evaluating large language model (LLM) output and making systematic comparisons to human-generated text. These methods can be used to characterize the psycholinguistic properties of LLM output and illustrate areas where LLMs fall short in comparison to human-generated text. In this work, we apply psycholinguistic methods to evaluate individual sentences from long-form analogies about biochemical concepts. We compare analogies generated by human subjects enrolled in introductory biochemistry courses to analogies generated by chatGPT. We perform a supervised classification analysis using 78 features extracted from Coh-metrix that analyze text cohesion, language, and readability (Graesser et. al., 2004). Results illustrate high performance for classifying student-generated and chatGPT-generated analogies. To evaluate which features contribute most to model performance, we use a hierarchical clustering approach. Results from this analysis illustrate several linguistic differences between the two sources.

arXiv:2306.03337

Form Follows Function: A Method of Discovering How Neurons Are Connected To Process Information

A method of discovering how neurons are connected to process information is presented here: Design a simple logic circuit that can perform a single, biologically advantageous function. Engineering concepts can be helpful in choosing the function and in designing the logic circuit. Several implementations of the method are reviewed to demonstrate how a biologically advantageous function can be chosen, how one simple network can generate major phenomena that are widely considered unrelated, and how one network design can lead to others that explain entirely different aspects of the brain. These results show that the method can benefit neuromorphic engineering as well as neuroscience, and that some brain functions can be carried out remarkably simply, at least in principle if not in the details.

arXiv:2306.02893

Branching model with state dependent offspring distribution for Chlamydia spread

Chlamydiae are bacteria with an interesting unusual developmental cycle. A single bacterium in its infectious form (elementary body, EB) enters the host cell, where it converts into its dividing form (reticulate body, RB), and divides by binary fission. Since only the EB form is infectious, before the host cell dies, RBs start to convert into EBs. After the host cell dies RBs do not survive. We model the population growth by a 2-type discrete-time branching process, where the probability of duplication depends on the state. Maximizing the EB production leads to a stochastic optimization problem. Simulation study shows that our novel model is able to reproduce the main features of the development of the population.

arXiv:2306.01746

An Application of Neutrosophic Sets to Decision Making

Maji et al. introduced in 2002 a method of parametric decision making using soft sets as tools and representing their tabular form as a binary matrix. In cases, however, where some or all of the parameters used for the characterization of the elements of the universal set are of fuzzy texture, their method does not give always the best decision making solution. In order to tackle this problem, we modified in earlier works the method of Maji et al. by replacing the binary elements in the tabular form of the corresponding soft set either by grey numbers or by triangular fuzzy numbers. In this work, in order to tackle more efficiently cases in which the decision maker has doubts about the correctness of the fuzzy/qualitative characterizations assigned to some or all of the elements of the universal set, we replace the binary elements of the tabular form by neutrosophic triplets. Our new, neutrosophic decision making method is illustrated by an application concerning the choice of a new player by a soccer club.

arXiv:2306.00740공식 코드 · Papers with Code

On the Limitations of Temperature Scaling for Distributions with Overlaps

Despite the impressive generalization capabilities of deep neural networks, they have been repeatedly shown to be overconfident when they are wrong. Fixing this issue is known as model calibration, and has consequently received much attention in the form of modified training schemes and post-training calibration procedures such as temperature scaling. While temperature scaling is frequently used because of its simplicity, it is often outperformed by modified training schemes. In this work, we identify a specific bottleneck for the performance of temperature scaling. We show that for empirical risk minimizers for a general set of distributions in which the supports of classes have overlaps, the performance of temperature scaling degrades with the amount of overlap between classes, and asymptotically becomes no better than random when there are a large number of classes. On the other hand, we prove that optimizing a modified form of the empirical risk induced by the Mixup data augmentation technique can in fact lead to reasonably good calibration performance, showing that training-time calibration may be necessary in some situations. We also verify that our theoretical results reflect practice by showing that Mixup significantly outperforms empirical risk minimization (with respect to multiple calibration metrics) on image classification benchmarks with class overlaps introduced in the form of label noise.

arXiv:2305.19271공식 코드 · Papers with Code

Concise Answers to Complex Questions: Summarization of Long-form Answers

Long-form question answering systems provide rich information by presenting paragraph-level answers, often containing optional background or auxiliary information. While such comprehensive answers are helpful, not all information is required to answer the question (e.g. users with domain knowledge do not need an explanation of background). Can we provide a concise version of the answer by summarizing it, while still addressing the question? We conduct a user study on summarized answers generated from state-of-the-art models and our newly proposed extract-and-decontextualize approach. We find a large proportion of long-form answers (over 90%) in the ELI5 domain can be adequately summarized by at least one system, while complex and implicit answers are challenging to compress. We observe that decontextualization improves the quality of the extractive summary, exemplifying its potential in the summarization task. To promote future work, we provide an extractive summarization dataset covering 1K long-form answers and our user study annotations. Together, we present the first study on summarizing long-form answers, taking a step forward for QA agents that can provide answers at multiple granularities.

arXiv:2305.18419

Semantic Segmentation with Bidirectional Language Models Improves Long-form ASR

We propose a method of segmenting long-form speech by separating semantically complete sentences within the utterance. This prevents the ASR decoder from needlessly processing faraway context while also preventing it from missing relevant context within the current sentence. Semantically complete sentence boundaries are typically demarcated by punctuation in written text; but unfortunately, spoken real-world utterances rarely contain punctuation. We address this limitation by distilling punctuation knowledge from a bidirectional teacher language model (LM) trained on written, punctuated text. We compare our segmenter, which is distilled from the LM teacher, against a segmenter distilled from a acoustic-pause-based teacher used in other works, on a streaming ASR pipeline. The pipeline with our segmenter achieves a 3.2% relative WER gain along with a 60 ms median end-of-segment latency reduction on a YouTube captioning task.

arXiv:2305.18201공식 코드 · Papers with Code

A Critical Evaluation of Evaluations for Long-form Question Answering

Long-form question answering (LFQA) enables answering a wide range of questions, but its flexibility poses enormous challenges for evaluation. We perform the first targeted study of the evaluation of long-form answers, covering both human and automatic evaluation practices. We hire domain experts in seven areas to provide preference judgments over pairs of answers, along with free-form justifications for their choices. We present a careful analysis of experts' evaluation, which focuses on new aspects such as the comprehensiveness of the answer. Next, we examine automatic text generation metrics, finding that no existing metrics are predictive of human preference judgments. However, some metrics correlate with fine-grained aspects of answers (e.g., coherence). We encourage future work to move away from a single "overall score" of the answer and adopt a multi-faceted evaluation, targeting aspects such as factuality and completeness. We publicly release all of our annotations and code to spur future work into LFQA evaluation.

arXiv:2305.16845공식 코드 · Papers with Code

An end-to-end strategy for recovering a free-form potential from a snapshot of stellar coordinates

New large observational surveys such as Gaia are leading us into an era of data abundance, offering unprecedented opportunities to discover new physical laws through the power of machine learning. Here we present an end-to-end strategy for recovering a free-form analytical potential from a mere snapshot of stellar positions and velocities. First we show how auto-differentiation can be used to capture an agnostic map of the gravitational potential and its underlying dark matter distribution in the form of a neural network. However, in the context of physics, neural networks are both a plague and a blessing as they are extremely flexible for modeling physical systems but largely consist in non-interpretable black boxes. Therefore, in addition, we show how a complementary symbolic regression approach can be used to open up this neural network into a physically meaningful expression. We demonstrate our strategy by recovering the potential of a toy isochrone system.

arXiv:2305.14604

Automated Market Making and Arbitrage Profits in the Presence of Fees

We consider the impact of trading fees on the profits of arbitrageurs trading against an automated marker marker (AMM) or, equivalently, on the adverse selection incurred by liquidity providers due to arbitrage. We extend the model of Milionis et al. [2022] for a general class of two asset AMMs to both introduce fees and discrete Poisson block generation times. In our setting, we are able to compute the expected instantaneous rate of arbitrage profit in closed form. When the fees are low, in the fast block asymptotic regime, the impact of fees takes a particularly simple form: fees simply scale down arbitrage profits by the fraction of time that an arriving arbitrageur finds a profitable trade.

arXiv:2305.14251공식 코드 · Papers with Code

FActScore: Fine-grained Atomic Evaluation of Factual Precision in Long Form Text Generation

Evaluating the factuality of long-form text generated by large language models (LMs) is non-trivial because (1) generations often contain a mixture of supported and unsupported pieces of information, making binary judgments of quality inadequate, and (2) human evaluation is time-consuming and costly. In this paper, we introduce FACTSCORE, a new evaluation that breaks a generation into a series of atomic facts and computes the percentage of atomic facts supported by a reliable knowledge source. We conduct an extensive human evaluation to obtain FACTSCOREs of people biographies generated by several state-of-the-art commercial LMs -- InstructGPT, ChatGPT, and the retrieval-augmented PerplexityAI -- and report new analysis demonstrating the need for such a fine-grained score (e.g., ChatGPT only achieves 58%). Since human evaluation is costly, we also introduce an automated model that estimates FACTSCORE using retrieval and a strong language model, with less than a 2% error rate. Finally, we use this automated metric to evaluate 6,500 generations from a new set of 13 recent LMs that would have cost $26K if evaluated by humans, with various findings: GPT-4 and ChatGPT are more factual than public models, and Vicuna and Alpaca are some of the best public models. FACTSCORE is available for public use via `pip install factscore`.

arXiv:2305.11353

Meta-learning for heterogeneous treatment effect estimation with closed-form solvers

This article proposes a meta-learning method for estimating the conditional average treatment effect (CATE) from a few observational data. The proposed method learns how to estimate CATEs from multiple tasks and uses the knowledge for unseen tasks. In the proposed method, based on the meta-learner framework, we decompose the CATE estimation problem into sub-problems. For each sub-problem, we formulate our estimation models using neural networks with task-shared and task-specific parameters. With our formulation, we can obtain optimal task-specific parameters in a closed form that are differentiable with respect to task-shared parameters, making it possible to perform effective meta-learning. The task-shared parameters are trained such that the expected CATE estimation performance in few-shot settings is improved by minimizing the difference between a CATE estimated with a large amount of data and one estimated with just a few data. Our experimental results demonstrate that our method outperforms the existing meta-learning approaches and CATE estimation methods.

arXiv:2305.11662공식 코드 · Papers with Code

Separating form and meaning: Using self-consistency to quantify task understanding across multiple senses

At the staggering pace with which the capabilities of large language models (LLMs) are increasing, creating future-proof evaluation sets to assess their understanding becomes more and more challenging. In this paper, we propose a novel paradigm for evaluating LLMs which leverages the idea that correct world understanding should be consistent across different (Fregean) senses of the same meaning. Accordingly, we measure understanding not in terms of correctness but by evaluating consistency across multiple senses that are generated by the model itself. We showcase our approach by instantiating a test where the different senses are different languages, hence using multilingual self-consistency as a litmus test for the model's understanding and simultaneously addressing the important topic of multilinguality. Taking one of the latest versions of ChatGPT as our object of study, we evaluate multilingual consistency for two different tasks across three different languages. We show that its multilingual consistency is still lacking, and that its task and world understanding are thus not language-independent. As our approach does not require any static evaluation corpora in languages other than English, it can easily and cheaply be extended to different languages and tasks and could become an integral part of future benchmarking efforts.

arXiv:2305.11058

Late-Binding Scholarship in the Age of AI: Navigating Legal and Normative Challenges of a New Form of Knowledge Production

Artificial Intelligence (AI) is poised to enable a new leap in the creation of scholarly content. New forms of engagement with AI systems, such as collaborations with large language models like GPT-3, offer affordances that will change the nature of both the scholarly process and the artifacts it produces. This article articulates ways in which those artifacts can be written, distributed, read, organized, and stored that are more dynamic, and potentially more effective, than current academic practices. Specifically, rather than the current "early-binding" process (that is, one in which ideas are fully reduced to a final written form before they leave an author's desk), we propose that there are substantial benefits to a "late-binding" process, in which ideas are written dynamically at the moment of reading. In fact, the paradigm of "binding" knowledge may transition to a new model in which scholarship remains ever "unbound" and evolving. An alternative form for a scholarly work could be encapsulated via several key components: a text abstract of the work's core arguments; hyperlinks to a bibliography of relevant related work; novel data that had been collected and metadata describing those data; algorithms or processes necessary for analyzing those data; a reference to a particular AI model that would serve as a "renderer" of the canonical version of the text; and specified parameters that would allow for a precise, word-for-word reconstruction of the canonical version. Such a form would enable both the rendering of the canonical version, and also the possibility of dynamic AI reimaginings of the text in light of future findings, scholarship unknown to the original authors, alternative theories, and precise tailoring to specific audiences (e.g., children, adults, professionals, amateurs).

arXiv:2305.10648공식 코드 · Papers with Code

Adjusting Logit in Gaussian Form for Long-Tailed Visual Recognition

It is not uncommon that real-world data are distributed with a long tail. For such data, the learning of deep neural networks becomes challenging because it is hard to classify tail classes correctly. In the literature, several existing methods have addressed this problem by reducing classifier bias, provided that the features obtained with long-tailed data are representative enough. However, we find that training directly on long-tailed data leads to uneven embedding space. That is, the embedding space of head classes severely compresses that of tail classes, which is not conducive to subsequent classifier learning. This paper therefore studies the problem of long-tailed visual recognition from the perspective of feature level. We introduce feature augmentation to balance the embedding distribution. The features of different classes are perturbed with varying amplitudes in Gaussian form. Based on these perturbed features, two novel logit adjustment methods are proposed to improve model performance at a modest computational overhead. Subsequently, the distorted embedding spaces of all classes can be calibrated. In such balanced-distributed embedding spaces, the biased classifier can be eliminated by simply retraining the classifier with class-balanced sampling data. Extensive experiments conducted on benchmark datasets demonstrate the superior performance of the proposed method over the state-of-the-art ones. Source code is available at https://github.com/Keke921/GCLLoss.

arXiv:2305.09991

Angle-based formation stabilization and maneuvers in port-Hamiltonian form with bearing and velocity measurements

This paper proposes a port-Hamiltonian framework for angle-based formation stabilization and maneuvers using bearing and velocity measurements with an underlying triangulated Laman graph. The corresponding port-Hamiltonian controller is designed using virtual couplings on the errors of angle constraints in angle space and then the angle constraints and agent actuators are mapped by the constraint Jacobian, which can be applied to other formation constraints. In addition, due to the fact that the port-Hamiltonian model allows for complex and heterogeneous agent dynamics, our framework can be extended to networks with different agent dynamics and formation constraints. To avoid unavailable distance terms in the control law, an estimator is designed based on port-Hamiltonian theory and the property that energy is coordinate-free for different sensor modalities using bearing and velocity measurements, which permits our framework to inject damping for the formation maneuvers. Furthermore, several maneuvers are analyzed under both considerations of stabilization and transient performance. Simulations are performed to illustrate the effectiveness of the approach.