Classifier-free guidance (CFG) is a key technique for improving conditional generation in diffusion models, enabling more accurate control while enhancing sample quality. It is natural to extend this technique to video diffusion, which generates video conditioned on a variable number of context frames, collectively referred to as history. However, we find two key challenges to guiding with variable-length history: architectures that only support fixed-size conditioning, and the empirical observation that CFG-style history dropout performs poorly. To address this, we propose the Diffusion Forcing Transformer (DFoT), a video diffusion architecture and theoretically grounded training objective that jointly enable conditioning on a flexible number of history frames. We then introduce History Guidance, a family of guidance methods uniquely enabled by DFoT. We show that its simplest form, vanilla history guidance, already significantly improves video generation quality and temporal consistency. A more advanced method, history guidance across time and frequency further enhances motion dynamics, enables compositional generalization to out-of-distribution history, and can stably roll out ex
Camera-controlled video generation has made substantial progress, enabling generated videos to follow prescribed viewpoint trajectories. However, existing methods usually learn camera-specific conditioning through camera encoders, control branches, or attention and positional-encoding modifications, which often require post-training on large-scale camera-annotated videos. Training-free alternatives avoid such post-training, but often shift the cost to test-time optimization or extra denoising-time guidance. We propose Warp-as-History, a simple interface that turns camera-induced warps into camera-warped pseudo-history with target-frame positional alignment and visible-token selection. Given a target camera trajectory, we construct camera-warped pseudo-history from past observations and feed it through the model's visual-history pathway. Crucially, we align its positional encoding with the target frames being denoised and remove warped-history tokens without valid source observations. Without any training, architectural modification, or test-time optimization, this interface reveals a non-trivial zero-shot capability of a frozen video generation model to follow camera trajectories.
This preliminary paper presents an interactive Evidence-RAG workspace for editorial assessment of AI-assisted peer review in the Journal of Digital History. The workflow makes model recommendations easier to inspect by linking reviewer comments, paper evidence, retrieval traces, and reproducibility checks. The system does not replace editors or reviewers. It treats large language models as auditable assistants whose outputs must be checked by human scholars. We describe the current pipeline: paper conversion, semantic chunking, vector indexing, retrieval-augmented evidence assessment, and a lightweight editorial interface. This is a preliminary version of a full paper associated with the accepted presentation "Towards an Interactive Evidence-RAG Peer-Review Workspace for the Journal of Digital History" at the conference AI through History, History through AI, C2DH, University of Luxembourg, 15-16 June 2026. The work was submitted to the conference on 26 February 2026. We also report a first editor-annotation analysis of 80 saved decisions for the Claude-Qwen audit configuration. For Claude-Qwen, strict editor-confirmed accuracy is 70.0%, the correct-or-mostly-correct rate is 86.2%.
Machine learning (ML) is the science of credit assignment. It seeks to find patterns in observations that explain and predict the consequences of events and actions. This then helps to improve future performance. Minsky's so-called "fundamental credit assignment problem" (1963) surfaces in all sciences including physics (why is the world the way it is?) and history (which persons/ideas/actions have shaped society and civilisation?). Here I focus on the history of ML itself. Modern artificial intelligence (AI) is dominated by artificial neural networks (NNs) and deep learning, both of which are conceptually closer to the old field of cybernetics than what was traditionally called AI (e.g., expert systems and logic programming). A modern history of AI & ML must emphasize breakthroughs outside the scope of shallow AI text books. In particular, it must cover the mathematical foundations of today's NNs such as the chain rule (1676), the first NNs (circa 1800), the first practical AI (1914), the theory of AI and its limitations (1931-34), and the first working deep learning algorithms (1965-). From the perspective of 2025, I provide a timeline of the most significant events in the hi
Infinitesimals have seen ups and downs in their tumultuous history. In the 18th century, d'Alembert set the tone by describing infinitesimals as chimeras. Some adversaries of infinitesimals, including Moigno and Connes, picked up on the term. We highlight the work of Cauchy, Noël, Poisson and Riemann. We also chronicle reactions by Moigno, Lamarle and Cantor, and signal the start of a revival with Peano.
A formalism is proposed to describe entangled quantum histories, and their entanglement entropy. We define a history vector, living in a tensor space with basis elements corresponding to the allowed histories, i.e. histories with nonvanishing amplitudes. The amplitudes are the components of the history vector, and contain the dynamical information. Probabilities of measurement sequences, and resulting collapse, are given by generalized Born rules: they are all expressed by means of projections and scalar products involving the history vector. Entangled history states are introduced, and a history density matrix is defined in terms of ensembles of history vectors. The corresponding history entropies (and history entanglement entropies for composite systems) are explicitly computed in two examples taken from quantum computation circuits.
Multiverse scenarios are common place in contemporary high energy physics and cosmology, although many consider them simply as bold speculations. In any case there is nothing like a single theory or a unified model of the multiverse. Instead, there are innumerable theoretical proposals often reciprocally incompatible. What is common to all these scenarios is the postulated existence of many causally disjointed regions of space/time (if not completely separated space-times) and the consideration of the observable universe as atypical in a global perspective. This paper examines the history of modern cosmology focusing on the forerunners of current ideas, and shows how some fundamental issues tend to present themselves on increasingly deeper levels of physical knowledge.
Modern Education is not \textit{Modern} without AI. However, AI's complex nature makes understanding and fixing problems challenging. Research worldwide shows that a parent's income greatly influences a child's education. This led us to explore how AI, especially complex models, makes important decisions using Explainable AI tools. Our research uncovered many complexities linked to parental income and offered reasonable explanations for these decisions. However, we also found biases in AI that go against what we want from AI in education: clear transparency and equal access for everyone. These biases can impact families and children's schooling, highlighting the need for better AI solutions that offer fair opportunities to all. This chapter tries to shed light on the complex ways AI operates, especially concerning biases. These are the foundational steps towards better educational policies, which include using AI in ways that are more reliable, accountable, and beneficial for everyone involved.
Space Very Long Baseline Interferometry is a radio astronomy technique distinguished by a record-high angular resolution reaching single-digit microseconds of arc. The paper provides a brief account of the history of developments of this technique over the period 1960s-2020s.
The formulation of a measurement theory for relativistic quantum field theory (QFT) has recently been an active area of research. In contrast to the asymptotic measurement framework that was enshrined in QED, the new proposals aim to supply a measurement framework for measurements in local spacetime regions. This paper surveys episodes in the history of quantum theory that contemporary researchers have identified as precursors to their own work and discusses how they laid the groundwork for current approaches to local measurement theory for QFT.
Not only motivated by the fact that the publication of the GAFT first appeared 60 years ago in print we reconstruct its history and so show that it is no exaggeration to claim that it has appeared already 75 years ago!
This paper is centered on some historical aspects of nuclear masses, and their relations to major discoveries. Besides nuclear reactions and decays, the heart of mass measurements lies in mass spectrometry, the early history of which will be reviewed first. I shall then give a short history of the mass unit which has not always been defined as one twelfth of the carbon-12 mass. When combining inertial masses from mass spectrometry with energy differences obtained in reactions and decays, the conversion factor between the two is essential. The history of the evaluation of the nuclear masses (actually atomic masses) is only slightly younger than that of the mass measurements themselves. In their modern form, mass evaluations can be traced back to 1955. Prior to 1955, several tables were established, the oldest one in 1935.
It is convenient to describe a quantum system at all times by means of a "history operator" $C$, encoding measurements and unitary time evolution between measurements. These operators naturally arise when computing the probability of measurement sequences, and generalize the "sum over position histories " of the Feynman path-integral. As we argue in the present note, this description has some computational advantages over the usual state vector description, and may help to clarify some issues regarding nonlocality of quantum correlations and collapse. A measurement on a system described by $C$ modifies the history operator, $C \rightarrow PC$, where $P$ is the projector corresponding to the measurement. We refer to this modification as "history operator collapse". Thus $C$ keeps track of the succession of measurements on a system, and contains all histories compatible with the results of these measurements. The collapse modifies the history content of $C$, and therefore modifies also the past (relative to the measurement), but never in a way to violate causality. Probabilities of outcomes are obtained as $Tr(C^\dagger P C)/Tr(C^\dagger C)$. A similar formula yields probabilities fo
Astrobiology is the field of science devoted to searching for life elsewhere in the Universe. It is inherently interdisciplinary, integrating results from multiple fields of science, and in this respect has strong synergies with 'big history'. I argue that big history and astrobiology are both acting to widen human perspectives in intellectually and socially beneficial directions, especially by enhancing public awareness of cosmic and evolutionary worldviews. I will further argue that these perspectives have important implications for the social and political organisation of humanity, including the eventual political unification of our planet. Astrobiology and big history are also concerned with the future of humanity, and I will argue that this future will be culturally and intellectually enriched if it includes the exploration of the universe around us.
We present in this chapter (Chapter II) the history of ideas which lead up to the development of modern knot theory. We are more detailed when pre-XX century history is reported. With more recent times we are more selective, stressing developments related to Jones type invariants of links. In the Appendix, A.Przybyszewska translation of Preface to P.Heegaard Dissertation (1898) is given.
On a centennial timescale, solar activity was quantified based on records of instrumental sunspot observations. This article briefly discusses several aspects of the recent archival investigations of historical sunspot records in the 17th to 18th centuries. This article also reviews the recent updates for the active day fraction and positions of the reported sunspot groups of the Maunder Minimum to show their significance within the observational history. These archival investigations serve as base datasets for reconstructing solar activity.
Throughout the early history of Science the heliocentric world model was refused because it contradicted the thoughts of Aristotle and the medieval "Impetus" theory of movement. Even Galileo's sky observations did not lead to any acceptance of the heliocentric model, because scientists derived from Aristotle's physics that the earth was static. It was not until 1640 when Gassendi proved the principle of inertia on a moving galley, as well as Galileo's laws of free fall with a giant wheel that "impetus-physics" which incorporated naive beliefs about movement was abandoned. To understand the methods of calculating movement, novice students can be stitmulaed by using these initial experiments based on the dynamics of a ship movement and not rely on naïve approach. This can help novice students to attain a concept change from naïve to scientific conceptions about movement. Modern students who study movement were found to harbour many of the same naive beliefs as our scientific forefathers. The history behind the elimination of these naive beliefs is presented here in order to help encourage elimination of these naive beliefs. Several studies found these naive beliefs to be held by novi
Recent event of ousting Rohingyas from Rakhine State by the Tatmadaw provoked worldwide public-and-academic interest in history and social evolution of the Rohingyas, and this is to what the article is devoted. As the existing literature presents a debate over Who are the Rohingyas?, and How legitimate is their claim over Rakhine State?, the paper reinvestigates the issues using a qualitative research method. Compiling a detailed history, the paper finds that Rohingya community developed through historically complicated processes marked by invasions and counter-invasions. The paper argues many people entered Bengal from Arakan before British brought people into Rakhine state. The Rohingyas believe Rakhine State is their ancestral homeland and they developed a sense of Ethnic Nationalism. Their right over Rakhine State is as significant as other groups. The paper concludes that the UN must pursue solution to the crisis and the government should accept the Rohingyas as it did the land or territory.
The goal of this work is to detect and recognize sequences of letters signed using fingerspelling in British Sign Language (BSL). Previous fingerspelling recognition methods have not focused on BSL, which has a very different signing alphabet (e.g., two-handed instead of one-handed) to American Sign Language (ASL). They also use manual annotations for training. In contrast to previous methods, our method only uses weak annotations from subtitles for training. We localize potential instances of fingerspelling using a simple feature similarity method, then automatically annotate these instances by querying subtitle words and searching for corresponding mouthing cues from the signer. We propose a Transformer architecture adapted to this task, with a multiple-hypothesis CTC loss function to learn from alternative annotation possibilities. We employ a multi-stage training approach, where we make use of an initial version of our trained model to extend and enhance our training data before re-training again to achieve better performance. Through extensive evaluations, we verify our method for automatic annotation and our model architecture. Moreover, we provide a human expert annotated te
This short essay traces the conceptual history of micro- and macroscopicity in the context of physical science. By focusing on three distinct episodes spanning five centuries, we show the scientific and philosophical meanings of this antonym pair, despite never being far from "the small" and "the large," have been evolving as the frontier of science advances. We analyze the intellectual and material impetus for these movements, and conclude that this conceptual history reflects the changing interaction between the natural world and humankind.