I am a 10th grade student and I don't know much about geology but I'm learning about subduction online. It's always mentioned that the denser plate gets subducted under the lesser denser plate. I want to know why it's only the denser plate which gets subducted? I know its buoyancy will be less, but still it's floating over the mantle and when it collides with the lesser denser plate why is it the one which gets subducted, or what "forces" play the role here?
Once subduction starts, it tends to keep going because the sinking slab pulls the material behind it. This is "slab pull", which dominates over "ridge push" at a subduction zone. How subduction starts is an open research question. How subduction continues once started is more or less settled science, so this answer focuses on what happens after subduction has started.
The reason the mafic oceanic crust (which is indeed less dense than ultramafic mantle rock) sinks is because at enough depth, the temperature becomes high enough to cook out the portions of the sinking slab with a lower melting point. Basalt and related rock in oceanic crust is a mix of multiple chemicals, some with lower melting points, some with higher melting points. At enough depth, the materials with a lower melting point melts while the materials with a higher melting point remain solid.
The materials with a lower melting point also have a lower density than the materials with a higher melting point. The less dense liquid portion of the partial melt rises through the overlying material to form the volcanic arc oftentimes associated with a subduction zone. This is the key process by which much of the continental crust formed.
What's left behind in the solid portion of the partial melt is dense ultramafic rock, basically the same chemically as mantle material. As melting is endothermic, the partial melting process cools the remaining solid portion even further. (It was already a bit cooler than the surrounding upper mantle.) That makes it even more dense, denser than the surrounding mantle material, and hence it sinks.
It is true that due to their chemical composition, oceanic plates (made up largely of basalt) are denser than continental plates (made up largely of granite). As a consequence, when a continental and an oceanic plate meet and push against each other, the oceanic plate generally loses and has to go underneath.
This would explain why subducting plates end up under the continents, but not why they actually keep sinking into the mantle. This may be surprising at first because the oceanic plate that is subducting is made up of the same material as the mantle (by and large). But it is much colder than the mantle, and colder generally means denser (with water/ice being the notable exception). This is because the subducting plate has been at the surface for tens of millions of years, cooling in contact with the oceans and the surface. So it is denser, and therefore sinks down in the surrounding mantle. We believe that some subducting plates sink all the way to the bottom of the mantle before they finally achieve temperature equilibrium with the surrounding mantle.
@Jean-Marie Prival's remark has the right info: Density and gravity are the forces at play to initiate a subduction.
But there also is a process of cooling and so time dependent increase of density at play on the side of the oceanic crust. Have a look at the concept of a Wilson cycle: a continent breaks up, a mid ocean ridge (MOR) forms, the new material gets welded to the edge of the continental crust, and as the ocean spreads by newly formed warmer material erupting at the MOR, it cools at the other end, near the continent, gets denser by this, and sinks into a ductile (not liquid!) layer called the Asthenosphere. Thus, the depth of an ocean, the isostatic equilibrium, becomes a function of age.
The older continental crust is made of less dense material than the oceanic crust anyway, but also doesn't cool. It will never subduct. It is because of this unwillingness that the ocean floor, welded to the continent, will at some point become so heavy under the force of gravity that it tears off, and starts to sink into the ductile layer, dragging the rest of the ocean floor behind it. Depending on factors it may on its way down stay colder and denser than the surrounding material, adjust the angle of subduction, a part may rip off, or even get stuck or push under the continent. But assume that it goes all the way down into the lower mantle.
So, there is no "collision" between oceanic and continental crust (not "plate", tectonic plates can have both types of crust on them), it is a gradual process. Collision plays a role once all of the ocean floor has been subducted and now the once split continental edges meet each other again, the Wilson cycle has completed. Then subduction stops, continental crust does not subduct, but gets compressed and shortened and something like the rise of the Himalayas and Tibetian plateau happens.
A consequence is that there is a maximum age for oceanic crust (~200 million years). Other than continental crust it constantly gets recycled as long as plate tectonics keep going.