Search

Mainstream approaches to aligning large language models (LLMs) heavily rely on human preference data, particularly when models require periodic updates. The standard process for iterative alignment of LLMs involves collecting new human feedback for each update. However, the data collection process is costly and challenging to scale. To address this issue, we introduce the "TS-Align" framework, which fine-tunes a policy model using pairwise feedback data automatically mined from its outputs. This automatic mining process is efficiently accomplished through the collaboration between a large-scale teacher model and a small-scale student model. The policy fine-tuning process can be iteratively repeated using on-policy generations within our proposed teacher-student collaborative framework. Through extensive experiments, we demonstrate that our final aligned policy outperforms the base policy model with an average win rate of 69.7% across seven conversational or instruction-following datasets. Furthermore, we show that the ranking capability of the teacher is effectively distilled into the student through our pipeline, resulting in a small-scale yet effective reward model for policy model alignment.

With copper-substituted lead apatite below room temperature, we observe diamagnetic dc magnetization under magnetic field of 25 Oe with remarkable bifurcation between zero-field-cooling and field-cooling measurements, and under 200 Oe it changes to be paramagnetism. A glassy memory effect is found during cooling. Typical hysteresis loops for superconductors are detected below 250 K, along with an asymmetry between forward and backward sweep of magnetic field. Our experiment suggests at room temperature the Meissner effect is possibly present in this material., Comment: 7 pages, 4 figures

We present magnetostriction and thermal expansion measurements on multiferroic (Ni0.93Co0.07)3V2O8. The high field phase diagrams up to 33 T along the a, b and c directions are built. For H//a, as the magnetic field increases, two intermediate phases appear between the incommensurate phase and the paramagnetic phase at about 7 K, and then a magnetically induced phase appears above the paramagnetic phase. For H//b,thermal expansion measurement indicates a mutation in the spin lattice coupling of the high field phases. The interlaced phase boundary suggests a mixed state in the optical high field phase. For H//c, an intermediate phase between the commensurate phase and the incommensurate phase is detected. A nonlinear boundary between the intermediate phase and the low temperature incommensurate phase, and a clear boundary between the commensurate phase and the paramagnetic phase are found. These results indicate that doping Co2+ breaks the weak ferromagnetic moment of the commensurate phase, which exists in the parent compound Ni3V2O8 and (Ni0.9Co0.1)3V2O8. This nonlinear influence reflects complicated spin modulation in Ni3V2O8 by doping Co2+., Comment: 7 pages, 4 figures

Recent advancements in reference-free learned metrics for open-domain dialogue evaluation have been driven by the progress in pre-trained language models and the availability of dialogue data with high-quality human annotations. However, current studies predominantly concentrate on English dialogues, and the generalization of these metrics to other languages has not been fully examined. This is largely due to the absence of a multilingual dialogue evaluation benchmark. To address the issue, we introduce xDial-Eval, built on top of open-source English dialogue evaluation datasets. xDial-Eval includes 12 turn-level and 6 dialogue-level English datasets, comprising 14930 annotated turns and 8691 annotated dialogues respectively. The English dialogue data are extended to nine other languages with commercial machine translation systems. On xDial-Eval, we conduct comprehensive analyses of previous BERT-based metrics and the recently-emerged large language models. Lastly, we establish strong self-supervised and multilingual baselines. In terms of average Pearson correlations over all datasets and languages, the best baseline outperforms OpenAI's ChatGPT by absolute improvements of 6.5% and 4.6% at the turn and dialogue levels respectively, albeit with much fewer parameters. The data and code are publicly available at https://github.com/e0397123/xDial-Eval., Comment: Accepted to EMNLP-2023 Findings

The performance of iron-based superconductors in high magnetic fields plays an important role for their practical application. In this work, we measured the magnetostriction and magnetization of BaFe$_{1.908}$Ni$_{0.092}$As$_{2}$ single crystals using pulsed magnetic fields up to 60 T and static magnetic fields up to 33 T, respectively. A huge longitudinal magnetostriction (of the order of 10$ ^{-4} $) was observed in the direction of the twin boundaries. The magnetization measurements evidence a high critical-current density due to strong bulk pinning. By using magnetization data with an exponential flux-pinning model, we can reproduce the magnetostriction curves qualitatively. This result shows that the magnetostriction of BaFe$_{1.908}$Ni$_{0.092}$As$_{2}$ can be well explained by a flux-pinning-induced mechanism., Comment: 4 pages, 3 figures

The advent and fast development of neural networks have revolutionized the research on dialogue systems and subsequently have triggered various challenges regarding their automatic evaluation. Automatic evaluation of open-domain dialogue systems as an open challenge has been the center of the attention of many researchers. Despite the consistent efforts to improve automatic metrics' correlations with human evaluation, there have been very few attempts to assess their robustness over multiple domains and dimensions. Also, their focus is mainly on the English language. All of these challenges prompt the development of automatic evaluation metrics that are reliable in various domains, dimensions, and languages. This track in the 11th Dialogue System Technology Challenge (DSTC11) is part of the ongoing effort to promote robust and multilingual automatic evaluation metrics. This article describes the datasets and baselines provided to participants and discusses the submission and result details of the two proposed subtasks.

In this study, the sol-gel method synthesized the magnetic measurement and analysis of single-phase polycrystalline perovskite DyFe1-xCrxO3 (DFCO). The experimental data were fitted and calculated by a four-sublattice molecular field model. Unlike previous studies, we found that in DyFe1-xCrxO3, the spin of the A-site rare earth ion Dy3+ also changed simultaneously with the spin reorientation of the Fe3+/Cr3+ ions. The effective spin is defined as the projection of the A site's total spin on the B site's spin plane, and the curve of temperature changes is obtained after fitting. With this theory, a very accurate thermomagnetic curve is obtained by fitting. This is convincing and, at the same time, provides a reference for the development of spintronic devices in the future.

$L^1$ based optimization is widely used in image denoising, machine learning and related applications. One of the main features of such approach is that it naturally provide a sparse structure in the numerical solutions. In this paper, we study an $L^1$ based mixed DG method for second-order elliptic equations in the non-divergence form. The elliptic PDE in nondivergence form arises in the linearization of fully nonlinear PDEs. Due to the nature of the equations, classical finite element methods based on variational forms can not be employed directly. In this work, we propose a new optimization scheme coupling the classical DG framework with recently developed $L^1$ optimization technique. Convergence analysis in both energy norm and $L^{\infty}$ norm are obtained under weak regularity assumption. Such $L^1$ models are nondifferentiable and therefore invalidate traditional gradient methods. Therefore all existing gradient based solvers are no longer feasible under this setting. To overcome this difficulty, we characterize solutions of $L^1$ optimization as fixed-points of proximity equations and utilize matrix splitting technique to obtain a class of fixed-point proximity algorithms with convergence analysis. Various numerical examples are displayed to illustrate the numerical solution has sparse structure with careful choice of the bases of the finite dimensional spaces. Numerical examples in both smooth and nonsmooth settings are provided to validate the theoretical results.

A subset $B$ of the ring $\mathbb{Z}_n$ is referred to as a $\ell$-covering set if $\{ ab \pmod n | 0\leq a \leq \ell, b\in B\} = \mathbb{Z}_n$. We show that there exists a $\ell$-covering set of $\mathbb{Z}_n$ of size $O(\frac{n}{\ell}\log n)$ for all $n$ and $\ell$, and how to construct such a set. We also provide examples where any $\ell$-covering set must have a size of $\Omega(\frac{n}{\ell}\frac{\log n}{\log \log n})$. The proof employs a refined bound for the relative totient function obtained through sieve theory and the existence of a large divisor with a linear divisor sum. The result can be used to simplify a modular subset sum algorithm.

Sophisticated spin instruments require high-precision spin control. In this study, we accurately study the intrinsic magnetic properties of the strongly disordered system NdFe0.5Cr0.5O3 through molecular field models combined with ASD theory. The three constituent sub-magnetic phases of the system are separated, and their magnetization contributions are calculated separately. Fitting the angle of the A/B magnetic moment at a given temperature, the reorientation temperature point and temperature dependence of different magnetic phases are obtained. This research will provide a very good theoretical support for studying complex disordered systems and applying high-precision spin control and lay a foundation for the design of new functional materials.

Text discourse parsing weighs importantly in understanding information flow and argumentative structure in natural language, making it beneficial for downstream tasks. While previous work significantly improves the performance of RST discourse parsing, they are not readily applicable to practical use cases: (1) EDU segmentation is not integrated into most existing tree parsing frameworks, thus it is not straightforward to apply such models on newly-coming data. (2) Most parsers cannot be used in multilingual scenarios, because they are developed only in English. (3) Parsers trained from single-domain treebanks do not generalize well on out-of-domain inputs. In this work, we propose a document-level multilingual RST discourse parsing framework, which conducts EDU segmentation and discourse tree parsing jointly. Moreover, we propose a cross-translation augmentation strategy to enable the framework to support multilingual parsing and improve its domain generality. Experimental results show that our model achieves state-of-the-art performance on document-level multilingual RST parsing in all sub-tasks., Comment: Published in CODI@EMNLP 2021

Speech evaluation is an essential component in computer-assisted language learning (CALL). While speech evaluation on English has been popular, automatic speech scoring on low resource languages remains challenging. Work in this area has focused on monolingual specific designs and handcrafted features stemming from resource-rich languages like English. Such approaches are often difficult to generalize to other languages, especially if we also want to consider suprasegmental qualities such as rhythm. In this work, we examine three different languages that possess distinct rhythm patterns: English (stress-timed), Malay (syllable-timed), and Tamil (mora-timed). We exploit robust feature representations inspired by music processing and vector representation learning. Empirical validations show consistent gains for all three languages when predicting pronunciation, rhythm and intonation performance., Comment: Accepted at INTERSPEECH 2021

Summarizing conversations via neural approaches has been gaining research traction lately, yet it is still challenging to obtain practical solutions. Examples of such challenges include unstructured information exchange in dialogues, informal interactions between speakers, and dynamic role changes of speakers as the dialogue evolves. Many of such challenges result in complex coreference links. Therefore, in this work, we investigate different approaches to explicitly incorporate coreference information in neural abstractive dialogue summarization models to tackle the aforementioned challenges. Experimental results show that the proposed approaches achieve state-of-the-art performance, implying it is useful to utilize coreference information in dialogue summarization. Evaluation results on factual correctness suggest such coreference-aware models are better at tracing the information flow among interlocutors and associating accurate status/actions with the corresponding interlocutors and person mentions., Comment: Accepted for presentation at SIGDIAL-2021. Version2: add BART-Large results/fix typos