arXiv subjects

Yu Bai

Publications and source records attributed to Yu Bai.

At least 91 records · Page 5Linked to original sources

arXiv2022

Conformal Predictor for Improving Zero-shot Text Classification Efficiency

Pre-trained language models (PLMs) have been shown effective for zero-shot (0shot) text classification. 0shot models based on natural language inference (NLI) and next sentence prediction (NSP) employ cross-encoder architecture and infer by making a forward pass through the model for each label-text pair separately. This increases the computational cost to make inferences linearly in the number of labels. In this work, we improve the efficiency of such cross-encoder-based 0shot models by restricting the number of likely labels using another fast base classifier-based conformal predictor (CP) calibrated on samples labeled by the 0shot model. Since a CP generates prediction sets with coverage guarantees, it reduces the number of target labels without excluding the most probable label based on the 0shot model. We experiment with three intent and two topic classification datasets. With a suitable CP for each dataset, we reduce the average inference time for NLI- and NSP-based models by 25.6% and 22.2% respectively, without dropping performance below the predefined error rate of 1%.

cs.CL↗

arXiv2022

Learning Rationalizable Equilibria in Multiplayer Games

A natural goal in multiagent learning besides finding equilibria is to learn rationalizable behavior, where players learn to avoid iteratively dominated actions. However, even in the basic setting of multiplayer general-sum games, existing algorithms require a number of samples exponential in the number of players to learn rationalizable equilibria under bandit feedback. This paper develops the first line of efficient algorithms for learning rationalizable Coarse Correlated Equilibria (CCE) and Correlated Equilibria (CE) whose sample complexities are polynomial in all problem parameters including the number of players. To achieve this result, we also develop a new efficient algorithm for the simpler task of finding one rationalizable action profile (not necessarily an equilibrium), whose sample complexity substantially improves over the best existing results of Wu et al. (2021). Our algorithms incorporate several novel techniques to guarantee rationalizability and no (swap-)regret simultaneously, including a correlated exploration scheme and adaptive learning rates, which may be of independent interest. We complement our results with a sample complexity lower bound showing the sharpness of our guarantees.

cs.LG↗

arXiv2022

The Role of Coverage in Online Reinforcement Learning

Coverage conditions -- which assert that the data logging distribution adequately covers the state space -- play a fundamental role in determining the sample complexity of offline reinforcement learning. While such conditions might seem irrelevant to online reinforcement learning at first glance, we establish a new connection by showing -- somewhat surprisingly -- that the mere existence of a data distribution with good coverage can enable sample-efficient online RL. Concretely, we show that coverability -- that is, existence of a data distribution that satisfies a ubiquitous coverage condition called concentrability -- can be viewed as a structural property of the underlying MDP, and can be exploited by standard algorithms for sample-efficient exploration, even when the agent does not know said distribution. We complement this result by proving that several weaker notions of coverage, despite being sufficient for offline RL, are insufficient for online RL. We also show that existing complexity measures for online RL, including Bellman rank and Bellman-Eluder dimension, fail to optimally capture coverability, and propose a new complexity measure, the sequential extrapolation coefficient, to provide a unification.

cs.LG↗

arXiv2022

PSP: Pre-trained Soft Prompts for Few-Shot Abstractive Summarization

Few-shot abstractive summarization has become a challenging task in natural language generation. To support it, we designed a novel soft prompts architecture coupled with a prompt pre-training plus fine-tuning paradigm that is effective and tunes only extremely light parameters. The soft prompts include continuous input embeddings across an encoder and a decoder to fit the structure of the generation models. Importantly, a novel inner-prompt placed in the text is introduced to capture document-level information. The aim is to devote attention to understanding the document that better prompts the model to generate document-related content. The first step in the summarization procedure is to conduct prompt pre-training with self-supervised pseudo-data. This teaches the model basic summarizing capabilities. The model is then fine-tuned with few-shot examples. Experimental results on the CNN/DailyMail and XSum datasets show that our method, with only 0.1% of the parameters, outperforms full-model tuning where all model parameters are tuned. It also surpasses Prompt Tuning by a large margin and delivers competitive results against Prefix-Tuning with 3% of the parameters.

cs.CL↗

arXiv2022

A long-period pre-ELM system discovered from LAMOST medium-resolution survey

We present LAMOST~J041920.07+072545.4 (hereafter J0419), a close binary consisting of a bloated extremely low mass pre-white dwarf (pre-ELM WD) and a compact object with an orbital period of 0.607189~days. The large-amplitude ellipsoidal variations and the evident Balmer and He~I emission lines suggest a filled Roche lobe and ongoing mass transfer. No outburst events were detected in the 15 years of monitoring of J0419, indicating a very low mass transfer rate. The temperature of the pre-ELM, $T_\mathrm{eff} = 5793_{-133}^{+124}\,\rm K$, is cooler than the known ELMs, but hotter than most CV donors. Combining the mean density within the Roche lobe and the radius constrained from our SED fitting, we obtain the mass of the pre-ELM, $M_1 = 0.176\pm 0.014\,M_\odot$. The joint fitting of light and radial velocity curves yields an inclination angle of $i = 66.5_{-1.7}^{+1.4}$ degrees, corresponding to the compact object mass of $M_2 = 1.09\pm 0.05\,M_\odot$. The very bloated pre-ELM has a smaller surface gravity ($\log g = 3.9\pm 0.01$, $R_1 = 0.78 \pm 0.02\,R_\odot$) than the known ELMs or pre-ELMs. The temperature and the luminosity ($L_\mathrm{bol} = 0.62_{-0.10}^{+0.11}\,L_\odot$) of J0419 are close to the main sequence, which makes the selection of such systems through the HR diagram inefficient. Based on the evolutionary model, the relatively long period and small $\log g$ indicate that J0419 could be close to the "bifurcation period" in the orbit evolution, which makes J0419 to be a unique source to connect ELM/pre-ELM WD systems, wide binaries and cataclysmic variables.

astro-ph.SR↗

arXiv2022

A White Dwarf-Main Sequence Binary Unveiled by Time-Domain Observations from LAMOST and TESS

We report a single-lined white dwarf-main sequence binary system, LAMOST J172900.17+652952.8, which is discovered by LAMOST's medium resolution time-domain surveys. The radial velocity semi-amplitude and orbital period of the optical visible star are measured by using the Palomar 200-inch telescope follow-up observations and the light curves from TESS. Thus the mass function of the invisible candidate white dwarf is derived, $f(M_{\rm{2}}) = 0.120\,\pm\,0.003\,M_{\odot}$. The mass of the visible star is measured based on the spectral energy distribution fitting, $M_{\mathrm{1}}$ = $0.81^{+0.07}_{-0.06}\,M_{\odot}$. Hence, the mass of its invisible companion is $M_{\rm{2}}\,\gtrsim\,0.63\,M_{\odot}$. The companion ought to be a compact object rather than a main-sequence star owing to the mass ratio $q = M_{\rm{2}} / M_{\rm 1} \gtrsim 0.78$ and the single-lined spectra. The compact object is likely to be a white dwarf except for small inclination angle, $i\,\lesssim\,40^{\circ}$. By using the GALEX NUV flux, the effective temperature of the white dwarf candidate is constrained as $T_{\rm eff}^{\rm WD}\,\lesssim\,12000-13500$ K. It is difficult to detect white dwarfs which are outshone by their bright companions via single-epoch optical spectroscopic surveys. Therefore, the optical time-domain surveys can play an important role in unveiling invisible white dwarfs and other compact objects in binaries.

astro-ph.SR↗

arXiv2022

A dynamically discovered and characterized non-accreting neutron star -- M dwarf binary candidate

Optical time-domain surveys can unveil and characterize exciting but less-explored non-accreting and/or non-beaming neutron stars (NS) in binaries. Here we report the discovery of such a NS candidate using the LAMOST spectroscopic survey. The candidate, designated LAMOST J112306.9+400736 (hereafter J1123), is in a single-lined spectroscopic binary containing an optically visible M star. The star's large radial velocity variation and ellipsoidal variations indicate a relatively massive unseen companion. Utilizing follow-up spectroscopy from the Palomar 200-inch telescope and high-precision photometry from TESS, we measure a companion mass of $1.24_{-0.03}^{+0.03}~M_{\odot}$. Main-sequence stars with this mass are ruled out, leaving a NS or a massive white dwarf (WD). Although a massive WD cannot be ruled out, the lack of UV excess radiation from the companion supports the NS hypothesis. Deep radio observations with FAST yielded no detections of either pulsed or persistent emission. J1123 is not detected in numerous X-ray and gamma-ray surveys. These non-detections suggest that the NS candidate is not presently accreting and pulsing. Our work exemplifies the capability of discovering compact objects in non-accreting close binaries by synergizing the optical time-domain spectroscopy and high-cadence photometry.

astro-ph.SR↗

arXiv2022

Local Calibration: Metrics and Recalibration

Probabilistic classifiers output confidence scores along with their predictions, and these confidence scores should be calibrated, i.e., they should reflect the reliability of the prediction. Confidence scores that minimize standard metrics such as the expected calibration error (ECE) accurately measure the reliability on average across the entire population. However, it is in general impossible to measure the reliability of an individual prediction. In this work, we propose the local calibration error (LCE) to span the gap between average and individual reliability. For each individual prediction, the LCE measures the average reliability of a set of similar predictions, where similarity is quantified by a kernel function on a pretrained feature space and by a binning scheme over predicted model confidences. We show theoretically that the LCE can be estimated sample-efficiently from data, and empirically find that it reveals miscalibration modes that are more fine-grained than the ECE can detect. Our key result is a novel local recalibration method LoRe, to improve confidence scores for individual predictions and decrease the LCE. Experimentally, we show that our recalibration method produces more accurate confidence scores, which improves downstream fairness and decision making on classification tasks with both image and tabular data.

cs.LG↗

arXiv2022

J-PLUS: Support Vector Regression to Measure Stellar Parameters

Context. Stellar parameters are among the most important characteristics in studies of stars, which are based on atmosphere models in traditional methods. However, time cost and brightness limits restrain the efficiency of spectral observations. The J-PLUS is an observational campaign that aims to obtain photometry in 12 bands. Owing to its characteristics, J-PLUS data have become a valuable resource for studies of stars. Machine learning provides powerful tools to efficiently analyse large data sets, such as the one from J-PLUS, and enable us to expand the research domain to stellar parameters. Aims. The main goal of this study is to construct a SVR algorithm to estimate stellar parameters of the stars in the first data release of the J-PLUS observational campaign. Methods. The training data for the parameters regressions is featured with 12-waveband photometry from J-PLUS, and is cross-identified with spectrum-based catalogs. These catalogs are from the LAMOST, the APOGEE, and the SEGUE. We then label them with the stellar effective temperature, the surface gravity and the metallicity. Ten percent of the sample is held out to apply a blind test. We develop a new method, a multi-model approach in order to fully take into account the uncertainties of both the magnitudes and stellar parameters. The method utilizes more than two hundred models to apply the uncertainty analysis. Results. We present a catalog of 2,493,424 stars with the Root Mean Square Error of 160K in the effective temperature regression, 0.35 in the surface gravity regression and 0.25 in the metallicity regression. We also discuss the advantages of this multi-model approach and compare it to other machine-learning methods.

astro-ph.IM↗

arXiv2022

Policy Optimization for Markov Games: Unified Framework and Faster Convergence

This paper studies policy optimization algorithms for multi-agent reinforcement learning. We begin by proposing an algorithm framework for two-player zero-sum Markov Games in the full-information setting, where each iteration consists of a policy update step at each state using a certain matrix game algorithm, and a value update step with a certain learning rate. This framework unifies many existing and new policy optimization algorithms. We show that the state-wise average policy of this algorithm converges to an approximate Nash equilibrium (NE) of the game, as long as the matrix game algorithms achieve low weighted regret at each state, with respect to weights determined by the speed of the value updates. Next, we show that this framework instantiated with the Optimistic Follow-The-Regularized-Leader (OFTRL) algorithm at each state (and smooth value updates) can find an $\mathcal{\widetilde{O}}(T^{-5/6})$ approximate NE in $T$ iterations, and a similar algorithm with slightly modified value update rule achieves a faster $\mathcal{\widetilde{O}}(T^{-1})$ convergence rate. These improve over the current best $\mathcal{\widetilde{O}}(T^{-1/2})$ rate of symmetric policy optimization type algorithms. We also extend this algorithm to multi-player general-sum Markov Games and show an $\mathcal{\widetilde{O}}(T^{-3/4})$ convergence rate to Coarse Correlated Equilibria (CCE). Finally, we provide a numerical example to verify our theory and investigate the importance of smooth value updates, and find that using "eager" value updates instead (equivalent to the independent natural policy gradient algorithm) may significantly slow down the convergence, even on a simple game with $H=2$ layers.

cs.LG↗

arXiv2022

Efficient and Differentiable Conformal Prediction with General Function Classes

Quantifying the data uncertainty in learning tasks is often done by learning a prediction interval or prediction set of the label given the input. Two commonly desired properties for learned prediction sets are \emph{valid coverage} and \emph{good efficiency} (such as low length or low cardinality). Conformal prediction is a powerful technique for learning prediction sets with valid coverage, yet by default its conformalization step only learns a single parameter, and does not optimize the efficiency over more expressive function classes. In this paper, we propose a generalization of conformal prediction to multiple learnable parameters, by considering the constrained empirical risk minimization (ERM) problem of finding the most efficient prediction set subject to valid empirical coverage. This meta-algorithm generalizes existing conformal prediction algorithms, and we show that it achieves approximate valid population coverage and near-optimal efficiency within class, whenever the function class in the conformalization step is low-capacity in a certain sense. Next, this ERM problem is challenging to optimize as it involves a non-differentiable coverage constraint. We develop a gradient-based algorithm for it by approximating the original constrained ERM using differentiable surrogate losses and Lagrangians. Experiments show that our algorithm is able to learn valid prediction sets and improve the efficiency significantly over existing approaches in several applications such as prediction intervals with improved length, minimum-volume prediction sets for multi-output regression, and label prediction sets for image classification.

cs.LG↗

arXiv2022

Sample-Efficient Learning of Correlated Equilibria in Extensive-Form Games

Imperfect-Information Extensive-Form Games (IIEFGs) is a prevalent model for real-world games involving imperfect information and sequential plays. The Extensive-Form Correlated Equilibrium (EFCE) has been proposed as a natural solution concept for multi-player general-sum IIEFGs. However, existing algorithms for finding an EFCE require full feedback from the game, and it remains open how to efficiently learn the EFCE in the more challenging bandit feedback setting where the game can only be learned by observations from repeated playing. This paper presents the first sample-efficient algorithm for learning the EFCE from bandit feedback. We begin by proposing $K$-EFCE -- a more generalized definition that allows players to observe and deviate from the recommended actions for $K$ times. The $K$-EFCE includes the EFCE as a special case at $K=1$, and is an increasingly stricter notion of equilibrium as $K$ increases. We then design an uncoupled no-regret algorithm that finds an $\varepsilon$-approximate $K$-EFCE within $\widetilde{\mathcal{O}}(\max_{i}X_iA_i^{K}/\varepsilon^2)$ iterations in the full feedback setting, where $X_i$ and $A_i$ are the number of information sets and actions for the $i$-th player. Our algorithm works by minimizing a wide-range regret at each information set that takes into account all possible recommendation histories. Finally, we design a sample-based variant of our algorithm that learns an $\varepsilon$-approximate $K$-EFCE within $\widetilde{\mathcal{O}}(\max_{i}X_iA_i^{K+1}/\varepsilon^2)$ episodes of play in the bandit feedback setting. When specialized to $K=1$, this gives the first sample-efficient algorithm for learning EFCE from bandit feedback.

cs.LG↗

arXiv2022

Unifying Cross-lingual Summarization and Machine Translation with Compression Rate

Cross-Lingual Summarization (CLS) is a task that extracts important information from a source document and summarizes it into a summary in another language. It is a challenging task that requires a system to understand, summarize, and translate at the same time, making it highly related to Monolingual Summarization (MS) and Machine Translation (MT). In practice, the training resources for Machine Translation are far more than that for cross-lingual and monolingual summarization. Thus incorporating the Machine Translation corpus into CLS would be beneficial for its performance. However, the present work only leverages a simple multi-task framework to bring Machine Translation in, lacking deeper exploration. In this paper, we propose a novel task, Cross-lingual Summarization with Compression rate (CSC), to benefit Cross-Lingual Summarization by large-scale Machine Translation corpus. Through introducing compression rate, the information ratio between the source and the target text, we regard the MT task as a special CLS task with a compression rate of 100%. Hence they can be trained as a unified task, sharing knowledge more effectively. However, a huge gap exists between the MT task and the CLS task, where samples with compression rates between 30% and 90% are extremely rare. Hence, to bridge these two tasks smoothly, we propose an effective data augmentation method to produce document-summary pairs with different compression rates. The proposed method not only improves the performance of the CLS task, but also provides controllability to generate summaries in desired lengths. Experiments demonstrate that our method outperforms various strong baselines in three cross-lingual summarization datasets. We released our code and data at https://github.com/ybai-nlp/CLS_CR.

cs.CL↗

arXiv2022

When Can We Learn General-Sum Markov Games with a Large Number of Players Sample-Efficiently?

Multi-agent reinforcement learning has made substantial empirical progresses in solving games with a large number of players. However, theoretically, the best known sample complexity for finding a Nash equilibrium in general-sum games scales exponentially in the number of players due to the size of the joint action space, and there is a matching exponential lower bound. This paper investigates what learning goals admit better sample complexities in the setting of $m$-player general-sum Markov games with $H$ steps, $S$ states, and $A_i$ actions per player. First, we design algorithms for learning an $ε$-Coarse Correlated Equilibrium (CCE) in $\widetilde{\mathcal{O}}(H^5S\max_{i\le m} A_i / ε^2)$ episodes, and an $ε$-Correlated Equilibrium (CE) in $\widetilde{\mathcal{O}}(H^6S\max_{i\le m} A_i^2 / ε^2)$ episodes. This is the first line of results for learning CCE and CE with sample complexities polynomial in $\max_{i\le m} A_i$. Our algorithm for learning CE integrates an adversarial bandit subroutine which minimizes a weighted swap regret, along with several novel designs in the outer loop. Second, we consider the important special case of Markov Potential Games, and design an algorithm that learns an $ε$-approximate Nash equilibrium within $\widetilde{\mathcal{O}}(S\sum_{i\le m} A_i / ε^3)$ episodes (when only highlighting the dependence on $S$, $A_i$, and $ε$), which only depends linearly in $\sum_{i\le m} A_i$ and significantly improves over existing efficient algorithm in the $ε$ dependence. Overall, our results shed light on what equilibria or structural assumptions on the game may enable sample-efficient learning with many players.

cs.LG↗

arXiv2022

Analyzing Micro-Founded General Equilibrium Models with Many Agents using Deep Reinforcement Learning

Real economies can be modeled as a sequential imperfect-information game with many heterogeneous agents, such as consumers, firms, and governments. Dynamic general equilibrium (DGE) models are often used for macroeconomic analysis in this setting. However, finding general equilibria is challenging using existing theoretical or computational methods, especially when using microfoundations to model individual agents. Here, we show how to use deep multi-agent reinforcement learning (MARL) to find $ε$-meta-equilibria over agent types in microfounded DGE models. Whereas standard MARL fails to learn non-trivial solutions, our structured learning curricula enable stable convergence to meaningful solutions. Conceptually, our approach is more flexible and does not need unrealistic assumptions, e.g., continuous market clearing, that are commonly used for analytical tractability. Furthermore, our end-to-end GPU implementation enables fast real-time convergence with a large number of RL economic agents. We showcase our approach in open and closed real-business-cycle (RBC) models with 100 worker-consumers, 10 firms, and a social planner who taxes and redistributes. We validate the learned solutions are $ε$-meta-equilibria through best-response analyses, show that they align with economic intuitions, and show our approach can learn a spectrum of qualitatively distinct $ε$-meta-equilibria in open RBC models. As such, we show that hardware-accelerated MARL is a promising framework for modeling the complexity of economies based on microfoundations.

cs.GT↗

arXiv2022

Detecting and Monitoring Tidal Dissipation of Hot Jupiters in the Era of SiTian

Transit Timing Variation (TTV) of hot Jupiters provides direct observational evidence of planet tidal dissipation. Detecting tidal dissipation through TTV needs high precision transit timings and long timing baselines. In this work, we predict and discuss the potential scientific contribution of SiTian Survey in detecting and analyzing exoplanet TTV. We develop a tidal dissipation detection pipeline for SiTian Survey that aims at time-domain astronomy with 72 1-meter optical telescopes. The pipeline includes the modules of light curve deblending, transit timing obtaining, and TTV modeling. SiTian is capable to detect more than 25,000 exoplanets among which we expect $\sim$50 sources showing evidence of tidal dissipation. We present detection and analysis of tidal dissipating targets, based on simulated SiTian light curves of XO-3b and WASP-161b. The transit light curve modeling gives consistent results within 1$σ$ to input values of simulated light curves. Also, the parameter uncertainties predicted by Monte-Carlo Markov Chain are consistent with the distribution obtained from simulating and modeling the light curve 1000 times. The timing precision of SiTian observations is $\sim$ 0.5 minutes with one transit visit. We show that differences between TTV origins, e.g., tidal dissipation, apsidal precession, multiple planets, would be significant, considering the timing precision and baseline. The detection rate of tidal dissipating hot Jupiters would answer a crucial question of whether the planet migrates at an early formation stage or random stages due to perturbations, e.g., planet scattering, secular interaction. SiTian identified targets would be constructive given that the sample would extend tenfold.

astro-ph.EP↗

arXiv2022

Privacy protection based on mask template

Powerful recognition algorithms are widely used in the Internet or important medical systems, which poses a serious threat to personal privacy. Although the law provides for diversity protection, e.g. The General Data Protection Regulation (GDPR) in Europe and Articles 1032 to 1039 of the civil code in China. However, as an important privacy disclosure event, biometric data is often hidden, which is difficult for the owner to detect and trace to the source. Human biometrics generally exist in images. In order to avoid the disclosure of personal privacy, we should prevent unauthorized recognition algorithms from acquiring the real features of the original image.

cs.CR↗

arXiv2022

Application of Color Block Code in Image Scaling

Aiming at the high cost of embedding annotation watermark in a narrow small area and the information distortion caused by the change of annotation watermark image resolution, this paper proposes a color block code technology, which uses location information and color code to form recognizable graphics, which can not only simplify the annotation graphics, but also ensure the recognition efficiency. First, the constituent elements of color block code are designed, and then the coding and decoding method of color block code is proposed. Experiments show that color block code has high anti-scaling and anti-interference, and can be widely used in the labeling of small object surface and low resolution image.

cs.GR↗