Tracking experiments in dense biological tissues reveal a diversity of sources for local energy injection at the cell scale. The effect of cell motility has been largely studied, but much less is known about the effect of the observed volume fluctuations of individual cells. We devise a simple microscopic model of 'actively-deforming' particles where local fluctuations of the particle size constitute a unique source of motion. We demonstrate that collective motion can emerge under the sole influence of such active volume fluctuations. We interpret the onset of diffusive motion as a nonequilibrium first-order phase transition, which arises at a well-defined amplitude of selfdeformation. This behaviour contrasts with the glassy dynamics produced by self-propulsion, but resembles the mechanical response of soft solids under mechanical deformation. It thus constitutes the first example of active yielding transition.Active matter represents a class of nonequilibrium systems that is currently under intense scrutiny [1,2]. In contrast to externally driven systems (such as sheared materials), active matter is driven out of equilibrium at the scale of its microscopic constituents. Well-studied examples include biological tissues [3], bacterial suspensions [4] and active granular and colloidal particles [5][6][7][8].Epithelial tissues constitute a biologically relevant active system composed of densely packed eukaryotic cells [3,[9][10][11][12][13][14]. Such tissues display a surprisingly fast and collective dynamics, which would not take place under equilibrium conditions [10]. This dynamics has been ascribed to at least three distinct active processes [13]: (i) self-propulsion through cell motility such as crawling [15], (ii) self-deformation through protrusion and contraction [16][17][18], and (iii) cell division and apoptosis [14]. The vertex model for tissues [12,19,20] includes the first two of these active processes and predicts a continuous static transition from an arrested to a flowing state [21]. Another theoretical line of research is based on self-propelled particles [22] which display at high density a nonequilibrium glass transition [23,24] accompanied by a continuous increase of space and time correlations which diverge on approaching the arrested phase [25][26][27]. However, typical correlation lengthscales in tissues do not seem to diverge [9,11,28,29].To disentangle the dynamic consequences of the various sources of activity in tissues at large scale, we suggest to decompose the original complex problem into simpler ones, and to study particle-based models which only include a specific source of activity. This strategy was followed earlier for self-propulsion, but experiments are instead often modelled by complex models with many competing processes [18,29,30]. We argue that it is relevant to introduce a simplified model to analyse the effect of active particle deformation in a dense assembly of nonpropelled soft objects. Specifically, we model a dense system that is driven out of equilibrium locally th...