Lowest Rated Papers
Enhancing pre-service teachers’ projective agency for diverse and multilingual classrooms through a course on curriculum development
Darío Luis Banegas, Michael Budzenski, Fang Yang
International Multilingual Research Journal·2024·5 citations
No abstract available
No ratings yet
View paper →DINO-WM: World Models on Pre-trained Visual Features enable Zero-shot Planning
Gaoyue Zhou, Hengkai Pan, Yann LeCun, Lerrel Pinto
arXiv·2024
The ability to predict future outcomes given control actions is fundamental for physical reasoning. However, such predictive models, often called world models, remains challenging to learn and are typically developed for task-specific solutions with online policy learning. To unlock world models' true potential, we argue that they should 1) be trainable on offline, pre-collected trajectories, 2) support test-time behavior optimization, and 3) facilitate task-agnostic reasoning. To this end, we present DINO World Model (DINO-WM), a new method to model visual dynamics without reconstructing the visual world. DINO-WM leverages spatial patch features pre-trained with DINOv2, enabling it to learn from offline behavioral trajectories by predicting future patch features. This allows DINO-WM to achieve observational goals through action sequence optimization, facilitating task-agnostic planning by treating goal features as prediction targets. We demonstrate that DINO-WM achieves zero-shot behavioral solutions at test time on six environments without expert demonstrations, reward modeling, or pre-learned inverse models, outperforming prior state-of-the-art work across diverse task families such as arbitrarily configured mazes, push manipulation with varied object shapes, and multi-particle scenarios.
No ratings yet
View paper →Um drácula brasileiro: análise da transposição do gótico na tradução de Lúcio Cardoso | Belas Infiéis
Aline Camargo Fontana, Guilherme Vinicius Lunelli, Jaqueline Bohn Donada, João Guilherme Marques dos Santos, Laura da Costa Toledo
periodicos.unb.br
This paper aims to analyze the translation of Drácula: O Homem da Noite, by Lúcio Cardoso, made in 1943, in comparison to the 1897 source text, originally written by Bram Stoker. Cardoso’s translation has significant relevance for the field of Translation Studies, as it was the first to appear in Brazilian Portuguese and one of the first to attempt the challenging task of transporting the novel’s Gothic elements in translation. The theoretical basis consists of theories of the Gothic as a literary mode, based on Botting (1996), França (2017), Groom (2012), Hogle (2002) and Spooner (2007). The strategy for carrying out the analysis revolves around the comparison between the source text and Cardoso’s translation, observing how the translator works with the Gothic and which omissions of original excerpts are identified (or not) in his translation. Results obtained indicate that Cardoso’s translation is significantly different from the source text, to the point that it is not possible, at certain moments of the narrative, to find correspondences between the texts in aspects that make Dracula a Gothic novel, as mentioned by Hogle (2002). Keywords: Translation. Gothic. Transposition of Gothic. Dracula. Lúcio Cardoso.
No ratings yet
View paper →Attitudes towards feminism predict susceptibility to feminism‐related fake news
Gillian Murphy, Emma Murray, Doireann Gough
Applied Cognitive Psychology·2021·17 citations
<jats:title>Summary</jats:title><jats:p>False memories may be especially likely when one is exposed to misinformation that is consistent with one's beliefs. Here, we assessed whether feminist attitudes predict susceptibility to feminism‐related fake news. In Experiment 1 (<jats:italic>n =</jats:italic> 1537), the more negative participants' attitudes towards feminism, the more likely they were to report a false memory for a fabricated event that negatively reflected on the feminist movement, and vice‐versa. This effect was only evident for those who interpreted the event as expected (e.g., those who rated the event as bad for feminism). When the purpose of the study was revealed, feminist attitudes also predicted ability to identify the stories as fake. We replicated these findings in Experiment 2 (<jats:italic>n =</jats:italic> 786), using fake stories that were less ambiguous. This study suggests that individuals are more susceptible to false memories for fake news stories that are ideologically congruent, even after a warning.</jats:p>
No ratings yet
View paper →Energy and Policy Considerations for Deep Learning in NLP
Emma Strubell, Ananya Ganesh, Andrew McCallum
Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics·2019·1552 citations
No abstract available
No ratings yet
View paper →Development of an Unmanned Aerial Vehicle Control System Capable of Adapting to Changing Weather Conditions and Obstacles
Ignat Epifantsev, Kristina Epifantseva
2024 International Conference on Industrial Engineering, Applications and Manufacturing (ICIEAM)·2024·1 citations
No abstract available
No ratings yet
View paper →The Interaction of Bengali and Japanese Artistic Milieus in the First Half of the Twentieth Century (1901-1945): Rabindranath Tagore, Arai Kanpō, and Nandalal Bose
Shigemi Inaga
Japan Review·2009
In both India and Japan, the literature on twentieth-century art history has been elaborated within the framework of nation-building. Japan enjoyed independence during the first half of that century, while India endured colonial rule. However, the difference between polities did not prevent intellectuals from the two cultural spheres from engaging in intensive interactions. This essay focuses on Okakura Kakuzō (Tenshin), author of The Ideals of the East (1904), and the painters Yokoyama Taikan, Hishida Shunsō, and Arai Kanpō. Yokoyama and Hishida were invited to India through Okakura's agency, and Yokoyama subsequently recommended Arai for an expedition to India. Exploring their deeds in this essay, the author seeks to shed new light on these figures' relationships with Rabindranath Tagore, Abanindranath Tagore, and Nandalal Bose. Okakura and these Japanese painters provided technical and iconographic inspiration to Nandalal, and as they did so they were exposed to early twentieth-century India. Their engagement with modern India does not exclude ideological dimensions, and the author touches on those here, as well. Fitting into a project that has a reevaluation of Asian modernism as its ultimate objective, this essay locates these examples of mutual influence between Japan and Bengal within the larger context of Asian intellectual history in the first half of the twentieth century. 1901~02年にインドに初滞在を果たした岡倉覚三は、菱田春草、横山大観をベンガルに派遣した。かれらの帰国の後、1913年岡倉の死後、アジア人として初めてノーベル文学賞を獲得した詩人、ロビンドロナート・タゴールが、1916年に日本を訪れ、原富太郎の三渓園に寄寓する。インドの詩人は下村観山が謡曲を題材とした《弱法師》に感激し、その複製を所望したことから、荒井寛方との交友が芽生え、インドに招かれた寛方は、ノンドラル・ボースほかの現地の画家と交友を育む一方、瀧精一の斡旋も得て、淫田専太郎ほかとともに、アジャンター壁画の模写に従事する。この事業は、追って帰国後、荒井晩年の法隆寺金堂壁画の模写につながる軌跡を描く。本稿では、岡倉の衣鉢を継ぐ日本美術院と、瀧精一が編集長を務めた『國華』との利害を兼備した荒井寛方の位置づけを検討し、その古典作品模写事業を支えた仏教的価値観に迫るとともに、大戦期におけるインドと日本の国民主義さらには超国家思想と美術造形との関わりを、宗教図像の展開のなかに復元する。詩人タゴールが盲目の《弱法師》の心眼に映る西方浄土の太陽の姿に扱んだ教訓は、西欧列強の支配下にあったインドの現状といかに関わっていたのか。それはボース晩年の東洋哲学ヘの傾倒を解き明かす鍵となるのか。そうした論点に論及し、とりあえずの仮説を提起することが、本稿の目的となる。
No ratings yet
View paper →Diffusion Transformers with Representation Autoencoders
Boyang Zheng, Nanye Ma, Shengbang Tong, Saining Xie
arXiv·2025
Latent generative modeling, where a pretrained autoencoder maps pixels into a latent space for the diffusion process, has become the standard strategy for Diffusion Transformers (DiT); however, the autoencoder component has barely evolved. Most DiTs continue to rely on the original VAE encoder, which introduces several limitations: outdated backbones that compromise architectural simplicity, low-dimensional latent spaces that restrict information capacity, and weak representations that result from purely reconstruction-based training and ultimately limit generative quality. In this work, we explore replacing the VAE with pretrained representation encoders (e.g., DINO, SigLIP, MAE) paired with trained decoders, forming what we term Representation Autoencoders (RAEs). These models provide both high-quality reconstructions and semantically rich latent spaces, while allowing for a scalable transformer-based architecture. Since these latent spaces are typically high-dimensional, a key challenge is enabling diffusion transformers to operate effectively within them. We analyze the sources of this difficulty, propose theoretically motivated solutions, and validate them empirically. Our approach achieves faster convergence without auxiliary representation alignment losses. Using a DiT variant equipped with a lightweight, wide DDT head, we achieve strong image generation results on ImageNet: 1.51 FID at 256x256 (no guidance) and 1.13 at both 256x256 and 512x512 (with guidance). RAE offers clear advantages and should be the new default for diffusion transformer training.
No ratings yet
View paper →Nested Learning: The Illusion of Deep Learning Architecture
Ali Behrouz, Meisam Razaviyayn, Peilin Zhong, and Vahab Mirrokni
Google Research
Over the last decades, developing more powerful neural architectures and simultaneously designing optimization algorithms to effectively train them have been the core of research efforts to enhance the capability of machine learning models. Despite the recent progresses, particularly in developing Language Models (LMs), there are fundamental challenges and unanswered questions about how such models can continually learn/memorize, self-improve, and find effective solutions. In this paper, we present a new learning paradigm, called Nested Learning (NL), that coherently represents a machine learning model with a set of nested, multi-level, and/or parallel optimization problems, each of which with its own “context f low”. Through the lenses of NL, existing deep learning methods learns from data through compressing their own context f low, and in-context learning naturally emerges in large models. NL suggests a philosophy to design more expressive learning algorithms with more “levels”, resulting in higher-order in-context learning and potentially unlocking effective continual learning capabilities. In addition to its neuro-scientific motivation, we advocate for NL by presenting three core contributions: (1) Expressive Optimizers: We show that known gradient-based optimizers, such as Adam, SGD with Momentum, etc., are in fact associative memory modules that aim to compress the gradients’ information (by gradient descent). Building on this insight, we present other “more expressive" optimizers with deep memory and/or more powerful learning rules; (2) Self-Modifying Learning Module: Taking advantage of NL’s insights on learning algorithms, we present a sequence model that learns how to modify itself by learning its own update algorithm; and (3) Continuum Memory System: Wepresent a new formulation for memory system that generalizes the traditional viewpoint of “long-term/short-term memory”. Combining our self-modifying sequence model with the continuum memory system, we present a continual learning module, called Hope, showing promising results in language modeling, knowledge incorporation, and few-shot generalization tasks, continual learning, and long-context reasoning tasks. “We cannot solve our problems with the same thinking we used when we created them!"
No ratings yet
View paper →A Case for Library-Level k-Means Binning in Histogram Gradient-Boosted Trees
Asher Labovich
arXiv·2025
Modern Gradient Boosted Decision Trees (GBDTs) accelerate split finding with histogram-based binning, which reduces complexity from $O(N\log N)$ to $O(N)$ by aggregating gradients into fixed-size bins. However, the predominant quantile binning strategy - designed to distribute data points evenly among bins -- may overlook critical boundary values that could enhance predictive performance. In this work, we consider a novel approach that replaces quantile binning with a $k$-means discretizer initialized with quantile bins, and justify the swap with a proof showing how, for any $L$-Lipschitz function, k-means maximizes the worst-case explained variance of Y obtained when treating all values in a given bin as equivalent. We test this swap against quantile and uniform binning on 33 OpenML datasets plus synthetics that control for modality, skew, and bin budget. Across 18 regression datasets, k-means shows no statistically significant losses at the 5% level and wins in three cases-most strikingly a 55% MSE drop on one particularly skewed dataset-even though k-means' mean reciprocal rank (MRR) is slightly lower (0.65 vs 0.72). On the 15 classification datasets the two methods are statistically tied (MRR 0.70 vs 0.68) with gaps $\leq$0.2 pp. Synthetic experiments confirm consistently large MSE gains - typically >20% and rising to 90% as outlier magnitude increases or bin budget drops. We find that k-means keeps error on par with exhaustive (no-binning) splitting when extra cuts add little value, yet still recovers key split points that quantile overlooks. As such, we advocate for a built-in bin_method=k-means flag, especially in regression tasks and in tight-budget settings such as the 32-64-bin GPU regime - because it is a "safe default" with large upside, yet adds only a one-off, cacheable overhead ($\approx$ 3.5s per feature to bin 10M rows on one Apple M1 thread).
No ratings yet
View paper →