Several theories are in play for what may be causing these objects to be active. One is sublimation, which is what drives comets. Why would that be a candidate here, then? Turns out, a thin layer of regolith as shallow as 1 meter in depth can cause ice to be trapped for almost a billion years, only becoming exposed when a collision occurs. Perhaps small pockets of ice formed in shadowed regions of asteroids and were not melted away by the radiation from the closeness of the sun. Maybe instead we are witnessing some projectiles coming from a recent collision with another space object, or maybe an object spinning apart because of a large torque. The problem is, the asteroid belt is not like how it looks in the movies. It is mainly empty space with the average distance between objects clocking in at 600,000 miles. With 800,000 asteroids in the belt, that translates to a lot of real estate available. Therefore, collisions should be quite rare (Jewitt, Redd 31).
Electrostatic forces may also be at play. Turns out, solar radiation involves a bombardment of not only photons but also electrons and protons. As an object spins in space, surfaces are hit with the radiation and electrons, being of a smaller mass, travel away faster than protons. This causes a net charge to develop as the objects spins and the surface falls into the dark side. But as it spins towards the light again, protons come into play again and electrostatic forces can cause particles to rise. If enough of a charge is developed, then the dust can achieve escape velocity and away they go. But the math shows it may only work for smaller asteroids, plus the moon models this is based off of may be incomplete (Jewitt).
Thermal properties can also be at hand. Fracturing caused by extreme temperature changes as an object approaches the sun can cause particles to escape. Another possibility is liquid water escaping the surface (as opposed to sublimation, where it goes directly from a solid to a gas), taking particles with it, whether that water loss be driven by heat differences or by shock compressions from collisions (Ibid).